CN114550053A - A method, device, computer equipment and storage medium for determining responsibility for a traffic accident - Google Patents
A method, device, computer equipment and storage medium for determining responsibility for a traffic accidentDownload PDFInfo
- Publication number
- CN114550053A CN114550053ACN202210171452.0ACN202210171452ACN114550053ACN 114550053 ACN114550053 ACN 114550053ACN 202210171452 ACN202210171452 ACN 202210171452ACN 114550053 ACN114550053 ACN 114550053A
- Authority
- CN
- China
- Prior art keywords
- accident
- video
- driving
- traffic
- vehicle
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/70—Information retrieval; Database structures therefor; File system structures therefor of video data
- G06F16/78—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/787—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using geographical or spatial information, e.g. location
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/215—Motion-based segmentation
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30241—Trajectory
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Evolutionary Computation (AREA)
- Mathematical Physics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Tourism & Hospitality (AREA)
- Multimedia (AREA)
- Databases & Information Systems (AREA)
- Artificial Intelligence (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- Library & Information Science (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Animal Behavior & Ethology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Traffic Control Systems (AREA)
Abstract
Description
Translated fromChinese技术领域technical field
本申请属于人工智能技术领域，具体涉及一种交通事故定责的方法、装置、计算机设备及存储介质。The present application belongs to the technical field of artificial intelligence, and specifically relates to a method, device, computer equipment and storage medium for determining responsibility for a traffic accident.
背景技术Background technique
在交警处理交通事故时，交警人员通常需要调取监控视频进行调查，才能确定事故发生的原因以判定事故责任，传统模式做法是在现场查勘结束后，回到交警部门的监控中心，再通过人工调阅监控视频，并对监控视频进行查阅，以完成交通事故定责处理。When the traffic police deal with traffic accidents, the traffic police usually need to obtain surveillance video for investigation, in order to determine the cause of the accident and determine the responsibility of the accident. Access the surveillance video, and check the surveillance video to complete the traffic accident determination and handling.
近年来，随着智能手机的普及，VoLTE视频通话、基站定位技术等技术也已趋近成熟，出现了一些通过线上调阅监控视频完成定责处理的方式，一般需要交警人员向监控中心提供事故地理位置信息，然后由监控中心的工作人员根据事故地理位置查找附近监控视频，再将监控视频发送给交警人员，交警人员通过对监控视频进行人工查阅，得到事故详情。上述线上调阅监控视频处理的方式需要交警人员和监控中心的工作人员多次对接确认监控视频是否记录到对应的事故信息，且监控视频需要交警人员或监控中心的工作人员人工翻阅大量视频资料才能获得，耗时较长，过程繁琐，导致交通事故案件处理效率较低。In recent years, with the popularization of smartphones, VoLTE video calling, base station positioning technology and other technologies have also become mature, and there have been some ways to complete responsibility determination through online access to surveillance video. Generally, traffic police personnel are required to provide the monitoring center with information. The location information of the accident is obtained, and then the staff of the monitoring center will find the nearby surveillance video according to the location of the accident, and then send the surveillance video to the traffic police, who will obtain the accident details by manually reviewing the surveillance video. The above-mentioned method of online access to surveillance video processing requires the traffic police personnel and the staff of the monitoring center to connect multiple times to confirm whether the corresponding accident information is recorded in the surveillance video, and the surveillance video requires the traffic police personnel or the staff of the monitoring center to manually read a large amount of video data. It takes a long time and the process is cumbersome, resulting in low efficiency in handling traffic accident cases.
发明内容SUMMARY OF THE INVENTION
本申请实施例的目的在于提出一种交通事故定责方法、装置、计算机设备及存储介质，以解决现有的人工查阅监控视频对交通事故进行定责的方案存在的耗时较长、过程繁琐，导致处理效率较低的技术问题。The purpose of the embodiments of the present application is to propose a method, device, computer equipment and storage medium for determining responsibility for a traffic accident, so as to solve the long time-consuming and cumbersome process of existing solutions for manually reviewing surveillance video to determine responsibility for traffic accidents , leading to technical problems with lower processing efficiency.
为了解决上述技术问题，本申请实施例提供一种交通事故定责方法，采用了如下所述的技术方案：In order to solve the above-mentioned technical problems, the embodiment of the present application provides a method for determining responsibility for a traffic accident, which adopts the following technical solutions:
一种交通事故定责的方法，包括：A method of determining liability for a traffic accident, including:
接收事故定责指令，根据所述事故定责指令对事故的位置进行定位，得到事故位置信息；Receive the accident responsibility determination instruction, locate the accident location according to the accident responsibility determination instruction, and obtain accident location information;
获取事故发生时间，根据所述事故位置信息和所述事故发生时间查找所述事故对应的监控视频；Obtain the accident occurrence time, and search for the surveillance video corresponding to the accident according to the accident location information and the accident occurrence time;
对所述监控视频进行解析，以确定所述监控视频中的关键帧图像；Analyzing the surveillance video to determine key frame images in the surveillance video;
对所述关键帧图像进行内容识别，确定所述关键帧图像中的事故车辆，并获取所述事故车辆的行车信息；Perform content recognition on the key frame image, determine the accident vehicle in the key frame image, and obtain the driving information of the accident vehicle;
在所述监控视频中对所述事故车辆进行标注，并将标注后的监控视频导入预先训练好的行驶轨迹识别模型，得到所述事故车辆的行驶轨迹；Marking the accident vehicle in the monitoring video, and importing the marked monitoring video into a pre-trained driving trajectory recognition model to obtain the driving trajectory of the accident vehicle;
获取预先构建的交管法规知识图谱，并基于所述行驶轨迹、所述行车信息和所述交管法规知识图谱判定所述事故车辆的事故责任。Acquire a pre-built traffic control law knowledge map, and determine the accident responsibility of the accident vehicle based on the driving track, the driving information, and the traffic control law knowledge map.
进一步地，所述获取事故发生时间，根据所述事故位置信息和所述事故发生时间查找所述事故对应的监控视频的步骤，具体包括：Further, the step of obtaining the accident occurrence time and searching for the monitoring video corresponding to the accident according to the accident location information and the accident occurrence time specifically includes:
根据所述事故位置信息查找事故发生地附近预设范围内的监控摄像头；Find surveillance cameras within a preset range near the accident site according to the accident location information;
调用所述监控摄像头的监控影像，并在所述监控影像中截取所述事故发生时间前后预设时长范围内的影像片段，得到所述事故对应的监控视频。The monitoring image of the monitoring camera is called, and the video clips within a preset time range before and after the occurrence time of the accident are intercepted from the monitoring image, so as to obtain the monitoring video corresponding to the accident.
进一步地，所述对所述监控视频进行解析，以确定所述监控视频中的关键帧图像的步骤，具体包括：Further, the step of analyzing the surveillance video to determine key frame images in the surveillance video specifically includes:
按照预设的时间间隔对所述监控视频进行划分，得到若干个视频片段；Divide the surveillance video according to a preset time interval to obtain several video clips;
分别计算每一个所述视频片段中各个视频帧的光流量；Calculate the optical flow of each video frame in each of the video segments respectively;
比对每一个所述视频片段中各个视频帧的光流量，并将光流量最小值对应的视频帧作为所述监控视频的关键帧图像。Compare the optical flow of each video frame in each of the video clips, and use the video frame corresponding to the minimum optical flow as the key frame image of the monitoring video.
进一步地，所述对所述监控视频进行解析，以确定所述监控视频中的关键帧图像的步骤，具体包括：Further, the step of analyzing the surveillance video to determine key frame images in the surveillance video specifically includes:
计算所述监控视频中相邻两个视频帧的直方图数据和灰度图数据；Calculate the histogram data and grayscale data of two adjacent video frames in the surveillance video;
基于所述直方图数据和所述灰度图数据，计算相邻两个视频帧之间的加权欧式距离；Calculate the weighted Euclidean distance between two adjacent video frames based on the histogram data and the grayscale image data;
基于所述加权欧式距离确定所述监控视频的镜头转换边界；determining the shot transition boundary of the surveillance video based on the weighted Euclidean distance;
基于所述镜头转换边界确定所述监控视频中的关键帧图像。A key frame image in the surveillance video is determined based on the shot transition boundary.
进一步地，所述行驶轨迹识别模型包括输入层、卷积层和输出层，所述在所述监控视频中对所述事故车辆进行标注，并将标注后的监控视频导入预先训练好的行驶轨迹识别模型，得到所述事故车辆的行驶轨迹的步骤，具体包括：Further, the driving trajectory recognition model includes an input layer, a convolution layer and an output layer, the accident vehicle is marked in the monitoring video, and the marked monitoring video is imported into the pre-trained driving trajectory. The steps of identifying the model to obtain the driving trajectory of the accident vehicle specifically include:
通过所述行驶轨迹识别模型的输入层对所述标注后的监控视频进行特征提取，并对提取到的视频特征进行特征向量转化，得到初始特征向量；Feature extraction is performed on the labeled surveillance video through the input layer of the driving track recognition model, and feature vector transformation is performed on the extracted video features to obtain an initial feature vector;
通过所述行驶轨迹识别模型的卷积层对所述初始特征向量进行卷积操作，得到局部特征权值矩阵；Convolution operation is performed on the initial feature vector through the convolution layer of the driving trajectory recognition model to obtain a local feature weight matrix;
通过所述行驶轨迹识别模型的输出层对所述局部特征权值矩阵进行组合，得到所述事故车辆的行驶轨迹。The local feature weight matrix is combined through the output layer of the driving trajectory recognition model to obtain the driving trajectory of the accident vehicle.
进一步地，在所述行驶轨迹识别模型包括输入层、卷积层和输出层，所述在所述监控视频中对所述事故车辆进行标注，并将标注后的监控视频导入预先训练好的行驶轨迹识别模型，得到所述事故车辆的行驶轨迹的步骤之前，还包括：Further, the driving trajectory recognition model includes an input layer, a convolution layer and an output layer, the accident vehicle is marked in the monitoring video, and the marked monitoring video is imported into the pre-trained driving. The trajectory recognition model, before the step of obtaining the driving trajectory of the accident vehicle, further includes:
获取训练视频，并对所述训练视频中的行驶车辆进行轨迹标注，得到所述行驶车辆的轨迹标注结果；Acquiring a training video, and labeling the trajectory of the driving vehicle in the training video to obtain a trajectory labeling result of the driving vehicle;
将所述训练视频导入预设的初始识别模型，其中，所述初始识别模型包括输入层、卷积层和输出层；importing the training video into a preset initial recognition model, wherein the initial recognition model includes an input layer, a convolution layer and an output layer;
通过所述初始识别模型的输入层对所述训练视频进行特征提取，并对提取到的视频特征进行特征向量转化，得到训练特征向量；Feature extraction is performed on the training video through the input layer of the initial recognition model, and feature vector transformation is performed on the extracted video features to obtain a training feature vector;
通过所述初始识别模型的卷积层对所述训练特征向量进行卷积操作，得到训练特征权值矩阵；Perform a convolution operation on the training feature vector through the convolution layer of the initial recognition model to obtain a training feature weight matrix;
通过所述初始识别模型的的输出层对所述训练特征权值矩阵进行组合，得到所述行驶车辆的行驶轨迹识别结果；The training feature weight matrix is combined through the output layer of the initial recognition model to obtain the driving trajectory recognition result of the driving vehicle;
基于所述行驶轨迹识别结果与所述轨迹标注结果，使用反向传播算法进行拟合，获取识别误差；Based on the driving trajectory recognition result and the trajectory labeling result, a back-propagation algorithm is used to perform fitting to obtain the recognition error;
将所述识别误差与预设误差阈值进行比较，若所述识别误差大于预设误差阈值，则对所述初始识别模型进行迭代更新，直到所述识别误差小于等于预设误差阈值为止，得到所述行驶轨迹识别模型。The identification error is compared with a preset error threshold, and if the identification error is greater than the preset error threshold, the initial identification model is iteratively updated until the identification error is less than or equal to the preset error threshold, and the result is obtained. The driving trajectory recognition model described above.
进一步地，所述基于所述行驶轨迹、所述行车信息和所述交管法规知识图谱判定所述事故车辆的事故责任的步骤，具体包括：Further, the step of judging the accident responsibility of the accident vehicle based on the driving track, the driving information and the knowledge map of traffic regulations specifically includes:
基于所述行驶轨迹和所述行车信息构建所述事故车辆的关联实体；constructing an associated entity of the accident vehicle based on the driving trajectory and the driving information;
将所述关联实体导入所述交管法规知识图谱，输出所述事故车辆的事故责任判定结果。The associated entity is imported into the traffic regulation knowledge map, and the accident responsibility determination result of the accident vehicle is output.
为了解决上述技术问题，本申请实施例还提供一种交通事故定责装置，采用了如下所述的技术方案：In order to solve the above-mentioned technical problems, the embodiment of the present application also provides a traffic accident liability determination device, which adopts the following technical solutions:
一种交通事故定责的装置，包括：A device for determining responsibility for a traffic accident, comprising:
事故定位模块，用于接收事故定责指令，根据所述事故定责指令对事故的位置进行定位，得到事故位置信息；The accident locating module is used to receive the accident responsibility determination instruction, locate the accident location according to the accident responsibility determination instruction, and obtain accident location information;
信息查询模块，用于获取事故发生时间，根据所述事故位置信息和所述事故发生时间查找所述事故对应的监控视频；an information query module, configured to obtain the accident occurrence time, and search for the surveillance video corresponding to the accident according to the accident location information and the accident occurrence time;
视频解析模块，用于对所述监控视频进行解析，以确定所述监控视频中的关键帧图像；a video parsing module for parsing the surveillance video to determine key frame images in the surveillance video;
内容识别模块，用于对所述关键帧图像进行内容识别，确定所述关键帧图像中的事故车辆，并获取所述事故车辆的行车信息；a content recognition module, configured to perform content recognition on the key frame image, determine the accident vehicle in the key frame image, and obtain the driving information of the accident vehicle;
轨迹识别模块，用于在所述监控视频中对所述事故车辆进行标注，并将标注后的监控视频导入预先训练好的行驶轨迹识别模型，得到所述事故车辆的行驶轨迹；a trajectory recognition module, used to mark the accident vehicle in the monitoring video, and import the marked monitoring video into a pre-trained driving trajectory recognition model to obtain the driving trajectory of the accident vehicle;
事故定责模块，用于获取预先构建的交管法规知识图谱，并基于所述行驶轨迹、所述行车信息和所述交管法规知识图谱判定所述事故车辆的事故责任。The accident responsibility determination module is used to obtain a pre-built traffic control law knowledge map, and determine the accident responsibility of the accident vehicle based on the driving track, the driving information and the traffic control law knowledge map.
为了解决上述技术问题，本申请实施例还提供一种计算机设备，采用了如下所述的技术方案：In order to solve the above-mentioned technical problems, the embodiment of the present application also provides a computer device, which adopts the following technical solutions:
一种计算机设备，包括存储器和处理器，所述存储器中存储有计算机可读指令，所述处理器执行所述计算机可读指令时实现如上述任一项所述的交通事故定责方法的步骤。A computer device, comprising a memory and a processor, wherein computer-readable instructions are stored in the memory, and when the processor executes the computer-readable instructions, the steps of the method for determining liability for a traffic accident as described in any of the above are implemented .
为了解决上述技术问题，本申请实施例还提供一种计算机可读存储介质，采用了如下所述的技术方案：In order to solve the above technical problems, the embodiments of the present application also provide a computer-readable storage medium, which adopts the following technical solutions:
一种计算机可读存储介质，所述计算机可读存储介质上存储有计算机可读指令，所述计算机可读指令被处理器执行时实现如上述中任一项所述的交通事故定责方法的步骤。A computer-readable storage medium on which computer-readable instructions are stored, and when the computer-readable instructions are executed by a processor, the method for determining liability for a traffic accident as described in any one of the above is realized. step.
与现有技术相比，本申请实施例主要有以下有益效果：Compared with the prior art, the embodiments of the present application mainly have the following beneficial effects:
本申请公开了一种交通事故定责的方法、装置、计算机设备及存储介质，属于人工智能技术领域。本申请通过交通事故的位置自动查找交通事故对应的监控视频，减少交警人员的工作量，然后对监控视频进行解析，以确定监控视频中的关键帧图像，对关键帧图像进行内容识别，确定交通事故中的事故车辆，然后在监控视频中对标注事故车辆，并将标注后的监控视频导入预先训练好的行驶轨迹识别模型，得到事故车辆的行驶轨迹，最后通过事故车辆的行驶轨迹、行车信息和预先构建的交管法规知识图谱自动判定事故车辆的事故责任。本申请通过事故定位和监控视频信息识别获取事故位置信息和事故车辆信息，并结合交管法规知识图谱实现对交通事故的自动定责，减轻交警人员的工作压力，提高交通事故案件处理效率。The present application discloses a method, device, computer equipment and storage medium for determining responsibility for a traffic accident, belonging to the technical field of artificial intelligence. This application automatically finds the surveillance video corresponding to the traffic accident through the location of the traffic accident, reduces the workload of the traffic police, and then parses the surveillance video to determine the key frame images in the surveillance video, performs content recognition on the key frame images, and determines the traffic The accident vehicle in the accident, then mark the accident vehicle in the surveillance video, and import the labeled surveillance video into the pre-trained driving trajectory recognition model to obtain the driving trajectory of the accident vehicle, and finally pass the driving trajectory and driving information of the accident vehicle. And the pre-built traffic control law knowledge map automatically determines the accident responsibility of the accident vehicle. The application obtains accident location information and accident vehicle information through accident location and monitoring video information identification, and combines the knowledge map of traffic control laws and regulations to realize automatic responsibility determination for traffic accidents, reduce the work pressure of traffic police personnel, and improve the efficiency of traffic accident case handling.
附图说明Description of drawings
为了更清楚地说明本申请中的方案，下面将对本申请实施例描述中所需要使用的附图作一个简单介绍，显而易见地，下面描述中的附图是本申请的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。In order to illustrate the solutions in the present application more clearly, the following will briefly introduce the accompanying drawings used in the description of the embodiments of the present application. For those of ordinary skill, other drawings can also be obtained from these drawings without any creative effort.
图1示出了本申请可以应用于其中的示例性系统架构图；FIG. 1 shows an exemplary system architecture diagram to which the present application can be applied;
图2示出了根据本申请的交通事故定责方法的一个实施例的流程图；FIG. 2 shows a flowchart of an embodiment of a method for determining responsibility for a traffic accident according to the present application;
图3示出了根据本申请的交通事故定责装置的一个实施例的结构示意图；FIG. 3 shows a schematic structural diagram of an embodiment of the device for determining responsibility for a traffic accident according to the present application;
图4示出了根据本申请的计算机设备的一个实施例的结构示意图。FIG. 4 shows a schematic structural diagram of an embodiment of a computer device according to the present application.
具体实施方式Detailed ways
除非另有定义，本文所使用的所有的技术和科学术语与属于本申请的技术领域的技术人员通常理解的含义相同；本文中在申请的说明书中所使用的术语只是为了描述具体的实施例的目的，不是旨在于限制本申请；本申请的说明书和权利要求书及上述附图说明中的术语“包括”和“具有”以及它们的任何变形，意图在于覆盖不排他的包含。本申请的说明书和权利要求书或上述附图中的术语“第一”、“第二”等是用于区别不同对象，而不是用于描述特定顺序。Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the technical field of this application; the terms used herein in the specification of the application are for the purpose of describing specific embodiments only It is not intended to limit the application; the terms "comprising" and "having" and any variations thereof in the description and claims of this application and the above description of the drawings are intended to cover non-exclusive inclusion. The terms "first", "second" and the like in the description and claims of the present application or the above drawings are used to distinguish different objects, rather than to describe a specific order.
在本文中提及“实施例”意味着，结合实施例描述的特定特征、结构或特性可以包含在本申请的至少一个实施例中。在说明书中的各个位置出现该短语并不一定均是指相同的实施例，也不是与其它实施例互斥的独立的或备选的实施例。本领域技术人员显式地和隐式地理解的是，本文所描述的实施例可以与其它实施例相结合。Reference herein to an "embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the present application. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor a separate or alternative embodiment that is mutually exclusive of other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.
为了使本技术领域的人员更好地理解本申请方案，下面将结合附图，对本申请实施例中的技术方案进行清楚、完整地描述。In order to make those skilled in the art better understand the solutions of the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the accompanying drawings.
如图1所示，系统架构100可以包括终端设备101、102、103，网络104和服务器105。网络104用以在终端设备101、102、103和服务器105之间提供通信链路的介质。网络104可以包括各种连接类型，例如有线、无线通信链路或者光纤电缆等等。As shown in FIG. 1 , the
用户可以使用终端设备101、102、103通过网络104与服务器105交互，以接收或发送消息等。终端设备101、102、103上可以安装有各种通讯客户端应用，例如网页浏览器应用、购物类应用、搜索类应用、即时通信工具、邮箱客户端、社交平台软件等。The user can use the
终端设备101、102、103可以是具有显示屏并且支持网页浏览的各种电子设备，包括但不限于智能手机、平板电脑、电子书阅读器、MP3播放器(Moving PictureExpertsGroup Audio Layer III，动态影像专家压缩标准音频层面3)、MP4(MovingPictureExperts Group Audio Layer IV，动态影像专家压缩标准音频层面4)播放器、膝上型便携计算机和台式计算机等等。The
服务器105可以是提供各种服务的服务器，例如对终端设备101、102、103上显示的页面提供支持的后台服务器，服务器可以是独立的服务器，也可以是提供云服务、云数据库、云计算、云函数、云存储、网络服务、云通信、中间件服务、域名服务、安全服务、内容分发网络(Content Delivery Network，CDN)、以及大数据和人工智能平台等基础云计算服务的云服务器。The
需要说明的是，本申请实施例所提供的交通事故定责方法一般由服务器执行，相应地，交通事故定责装置一般设置于服务器中。It should be noted that the method for determining responsibility for a traffic accident provided by the embodiments of the present application is generally executed by a server, and accordingly, a device for determining responsibility for a traffic accident is generally set in the server.
应该理解，图1中的终端设备、网络和服务器的数目仅仅是示意性的。根据实现需要，可以具有任意数目的终端设备、网络和服务器。It should be understood that the numbers of terminal devices, networks and servers in FIG. 1 are merely illustrative. There can be any number of terminal devices, networks and servers according to implementation needs.
继续参考图2，示出了根据本申请的交通事故定责方法的一个实施例的流程图。本申请实施例可以基于人工智能技术对相关的数据进行获取和处理。其中，人工智能(Artificial Intelligence，AI)是利用数字计算机或者数字计算机控制的机器模拟、延伸和扩展人的智能，感知环境、获取知识并使用知识获得最佳结果的理论、方法、技术及应用系统。Continuing to refer to FIG. 2 , a flow chart of an embodiment of the method for determining liability for a traffic accident according to the present application is shown. The embodiments of the present application may acquire and process related data based on artificial intelligence technology. Among them, artificial intelligence (AI) is a theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results. .
人工智能基础技术一般包括如传感器、专用人工智能芯片、云计算、分布式存储、大数据处理技术、操作/交互系统、机电一体化等技术。人工智能软件技术主要包括计算机视觉技术、机器人技术、生物识别技术、语音处理技术、自然语言处理技术以及机器学习/深度学习等几大方向。The basic technologies of artificial intelligence generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation/interaction systems, and mechatronics. Artificial intelligence software technology mainly includes computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning/deep learning.
近年来，随着智能手机的普及，VoLTE视频通话、基站定位技术等技术也已趋近成熟，出现了一些通过线上调阅监控视频完成定责处理的方式，这种处理方式一般需要交警人员向监控中心提供事故地理位置信息，然后由监控中心的工作人员根据事故地理位置查找附近监控视频，监控中心的工作人员通常需要耗费大量的时间来翻阅事故地理位置周围所有的监控视频，以确定记录有完整事故经过的监控视频，查阅过程繁琐，耗时较长，导致交通事故案件处理效率较低。In recent years, with the popularization of smartphones, VoLTE video calling, base station positioning technology and other technologies have also become mature, and there have been some ways to complete responsibility determination through online review of surveillance video, which generally requires traffic police officers. Provide the accident location information to the monitoring center, and then the monitoring center staff will find nearby surveillance videos based on the accident location. The monitoring center staff usually spend a lot of time reading all the surveillance videos around the accident location to determine the record. There is a complete monitoring video of the accident, and the review process is cumbersome and time-consuming, resulting in low efficiency in handling traffic accident cases.
为此，本申请公开一种交通事故定责的方法、装置、计算机设备及存储介质，旨在解决上述人工查阅监控视频对交通事故进行定责的方案存在的耗时较长、过程繁琐，导致处理效率较低的技术问题。所述的交通事故定责方法，包括以下步骤：To this end, the present application discloses a method, device, computer equipment and storage medium for determining responsibility for a traffic accident, which aims to solve the long time-consuming and cumbersome process of the above-mentioned solution of manually reviewing surveillance video to determine responsibility for a traffic accident, resulting in Deal with less efficient technical issues. The method for determining liability for a traffic accident includes the following steps:
S201，接收事故定责指令，根据所述事故定责指令对事故的位置进行定位，得到事故位置信息。S201: Receive an accident responsibility determination instruction, locate the accident location according to the accident responsibility determination instruction, and obtain accident location information.
具体的，服务器接收到交通事故定责指令后，根据指令对交通事故的位置进行定位，得到事故位置信息。例如，报案人在第一现场拨打电话报警时，本申请的报案系统自动建立报案人与交警的VoLTE视频通话，通过此方式，报案系统根据基站定位取得报案人的当前位置作为事故发生地点，获取事故发生地点的经纬度信息，且建立VoLTE视频通话方便交警远程查看事故现场情况，指引报案人远程查勘。Specifically, after receiving the traffic accident responsibility determination instruction, the server locates the location of the traffic accident according to the instruction, and obtains accident location information. For example, when the reporter calls the police at the first scene, the reporting system of the present application automatically establishes a VoLTE video call between the complainant and the traffic police. The longitude and latitude information of the accident location, and the establishment of a VoLTE video call facilitates the traffic police to remotely view the accident scene and guide the informant to investigate remotely.
在上述实施例中，交通事故定责方法运行于其上的电子设备(例如图1所示的服务器)可以通过有线连接方式或者无线连接方式接收事故定责指令。需要指出的是，上述无线连接方式可以包括但不限于3G/4G连接、WiFi连接、蓝牙连接、WiMAX连接、Zigbee连接、UWB(ultra wideband)连接、以及其他现在已知或将来开发的无线连接方式。In the above embodiment, the electronic device (for example, the server shown in FIG. 1 ) on which the method for determining responsibility for a traffic accident runs may receive an instruction for determining responsibility for an accident through a wired connection or a wireless connection. It should be pointed out that the above wireless connection methods may include but are not limited to 3G/4G connection, WiFi connection, Bluetooth connection, WiMAX connection, Zigbee connection, UWB (ultra wideband) connection, and other wireless connection methods currently known or developed in the future .
S202，获取事故发生时间，根据所述事故位置信息和所述事故发生时间查找所述事故对应的监控视频。S202: Acquire the accident occurrence time, and search for the surveillance video corresponding to the accident according to the accident location information and the accident occurrence time.
具体的，服务器根据事故位置信息查找事故发生地附近500m范围内的所有监控摄像头，调用所有监控摄像头在报案人发起事故定责指令前1个小时的监控影像，并通过场景识别的方式在监控影像中快速查找本次交通事故对应的监控视频。Specifically, the server searches for all surveillance cameras within 500m of the accident location according to the accident location information, calls the surveillance images of all surveillance cameras 1 hour before the reporter initiates the accident determination instruction, and uses scene recognition to display the surveillance images in the surveillance images. Quickly find the surveillance video corresponding to this traffic accident.
S203，对所述监控视频进行解析，以确定所述监控视频中的关键帧图像。S203: Analyze the surveillance video to determine key frame images in the surveillance video.
具体的，服务器对监控视频进行解析，并通过计算视频帧光流量或视频帧镜头转换边界的方法确定监控视频中的关键帧图像。其中，关键帧图像主要用于确定事故车辆，以及获取部分事故信息，如车辆碰撞情况等。Specifically, the server parses the surveillance video, and determines the key frame images in the surveillance video by calculating the optical flow of the video frame or the transition boundary of the video frame shot. Among them, the key frame image is mainly used to determine the accident vehicle, and to obtain some accident information, such as vehicle collision situation and so on.
S204，对所述关键帧图像进行内容识别，确定所述关键帧图像中的事故车辆，并获取所述事故车辆的行车信息。S204: Perform content recognition on the key frame image, determine the accident vehicle in the key frame image, and acquire the driving information of the accident vehicle.
具体的，服务器通过预先训练好的内容识别模型对关键帧图像进行内容识别，确定关键帧图像中的路面信息和车辆信息，并进一步识别关键帧图像中的事故车辆，然后在监控视频中对事故车辆进行追踪，从监控视频中分析事故车辆的行车信息，例如行车速度、行车车道等等。Specifically, the server performs content recognition on the key frame image through the pre-trained content recognition model, determines the road information and vehicle information in the key frame image, and further identifies the accident vehicle in the key frame image, and then identifies the accident in the monitoring video. The vehicle is tracked, and the driving information of the accident vehicle, such as driving speed, driving lane, etc., is analyzed from the surveillance video.
S205，在所述监控视频中对所述事故车辆进行标注，并将标注后的监控视频导入预先训练好的行驶轨迹识别模型，得到所述事故车辆的行驶轨迹。S205 , label the accident vehicle in the surveillance video, and import the labeled surveillance video into a pre-trained driving trajectory recognition model to obtain the driving trajectory of the accident vehicle.
其中，行驶轨迹识别模型用于车辆行驶轨迹的识别，行驶轨迹识别模型基于卷积神经网络结构进行搭建，卷积神经网络(Convolutional Neural Networks,CNN)是一类包含卷积计算且具有深度结构的前馈神经网络(Feedforward Neural Networks)，是深度学习(deep learning)的代表算法之一。卷积神经网络具有表征学习(representationlearning)能力，能够按其阶层结构对输入信息进行平移不变分类(shift-invariantclassification)，因此也被称为“平移不变人工神经网络”。卷积神经网络仿造生物的视知觉(visual perception)机制构建，可以进行监督学习和非监督学习，其卷积层内的卷积核参数共享和层间连接的稀疏性使得卷积神经网络能够以较小的计算量对格点化(grid-like topology)特征，例如像素和音频进行学习，有稳定的效果且对数据没有额外的特征工程要求。Among them, the driving trajectory recognition model is used to identify the driving trajectory of the vehicle, and the driving trajectory recognition model is constructed based on the convolutional neural network structure. Feedforward Neural Networks is one of the representative algorithms of deep learning. Convolutional neural network has the ability of representation learning and can perform shift-invariant classification of input information according to its hierarchical structure, so it is also called "shift-invariant artificial neural network". Convolutional neural networks are constructed by imitating the visual perception mechanism of biology, and can perform supervised learning and unsupervised learning. Small computational effort to learn grid-like topology features such as pixels and audio, with stable results and no additional feature engineering requirements on the data.
具体的，服务器预先基于卷积神经网络结构搭建行驶轨迹识别模型，在进行行驶轨迹识别时，通过在监控视频中对事故车辆进行标注，并将标注后的监控视频导入预先训练好的行驶轨迹识别模型，得到事故车辆的行驶轨迹。Specifically, the server builds a driving trajectory recognition model based on the convolutional neural network structure in advance. When performing driving trajectory recognition, it labels the accident vehicle in the monitoring video, and imports the marked monitoring video into the pre-trained driving trajectory recognition model. model to obtain the driving trajectory of the accident vehicle.
S206，获取预先构建的交管法规知识图谱，并基于所述行驶轨迹、所述行车信息和所述交管法规知识图谱判定所述事故车辆的事故责任。S206: Acquire a pre-built traffic control law knowledge map, and determine the accident responsibility of the accident vehicle based on the driving track, the driving information, and the traffic control law knowledge map.
其中，知识图谱作为一种语义网络拥有极强的表达能力和建模灵活性。首先，知识图谱是一种语义表示，可以对现实世界中的实体、概念、属性以及它们之间的关系进行建模；其次，知识图谱是其衍生技术的数据交换标准，其本身是一种数据建模的“协议”，相关技术涵盖知识抽取、知识集成、知识管理和知识应用等各个环节。在本申请中，通过预先获取历史交通事故案例构建交管法规知识图谱，通过交管法规知识图谱识别事故车辆的责任。Among them, knowledge graph, as a kind of semantic network, has strong expressive ability and modeling flexibility. First, the knowledge graph is a semantic representation that can model entities, concepts, attributes and their relationships in the real world; secondly, the knowledge graph is a data exchange standard for its derived technologies, which itself is a kind of data The "protocol" of modeling, related technologies cover all aspects of knowledge extraction, knowledge integration, knowledge management and knowledge application. In this application, a knowledge map of traffic regulations is constructed by obtaining historical traffic accident cases in advance, and the responsibility of the accident vehicle is identified through the knowledge map of traffic regulations.
具体的，服务器获取预先构建的交管法规知识图谱，并基于事故车辆的行驶轨迹、事故车辆的行车信息构建事故车辆的关联实体，根据构建的事故车辆的关联实体和交管法规知识图谱判定事故车辆的事故责任。Specifically, the server obtains a pre-built knowledge map of traffic control regulations, and builds an associated entity of the accident vehicle based on the driving track of the accident vehicle and the driving information of the accident vehicle, and determines the accident vehicle according to the constructed associated entity of the accident vehicle and the traffic control law knowledge map. accident liability.
在上述实施例中，本申请通过事故定位和监控视频信息识别获取事故位置信息和事故车辆信息，并结合交管法规知识图谱实现对交通事故的自动定责，减轻交警人员的工作压力，提高交通事故案件处理效率。In the above-mentioned embodiment, the present application obtains accident location information and accident vehicle information through accident location and monitoring video information identification, and realizes automatic responsibility determination for traffic accidents in combination with the knowledge map of traffic control regulations, reduces the work pressure of traffic police personnel, and improves traffic accidents. Case handling efficiency.
进一步地，所述获取事故发生时间，根据所述事故位置信息和所述事故发生时间查找所述事故对应的监控视频的步骤，具体包括：Further, the step of obtaining the accident occurrence time and searching for the monitoring video corresponding to the accident according to the accident location information and the accident occurrence time specifically includes:
根据所述事故位置信息查找事故发生地附近预设范围内的监控摄像头；Find surveillance cameras within a preset range near the accident site according to the accident location information;
调用所述监控摄像头的监控影像，并在所述监控影像中截取所述事故发生时间前后预设时长范围内的影像片段，得到所述事故对应的监控视频。The monitoring image of the monitoring camera is called, and the video clips within a preset time range before and after the occurrence time of the accident are intercepted from the monitoring image, so as to obtain the monitoring video corresponding to the accident.
具体的，服务器根据事故位置信息查找事故发生地附近500m范围内的所有监控摄像头，调用事故发生地附近500m范围内所有监控摄像头在报案人发起事故定责指令前1个小时和后1个小时内的监控影像，并通过场景识别的方式在监控影像中快速查找本次交通事故对应的监控视频。其中，每一个连接系统的监控摄像头都会预先上传一张拍摄到的场景图像，以构建场景图像库。服务器在接收事故定责指令之后，会指示报案人上传一张事故现场附近的照片，服务器通过比对事故现场附近的照片和场景图像库中的场景图像之间的相似度，确定拍摄到事故发生的监控摄像头，并提取该监摄像头的监控视频。Specifically, the server searches for all surveillance cameras within 500m of the accident location according to the accident location information, and calls all surveillance cameras within 500 meters of the accident location within 1 hour before and 1 hour after the reporter initiates the accident determination instruction and quickly find the surveillance video corresponding to the traffic accident in the surveillance image by means of scene recognition. Among them, each surveillance camera connected to the system will upload a captured scene image in advance to build a scene image library. After receiving the accident responsibility instruction, the server will instruct the reporter to upload a photo near the accident scene. surveillance camera, and extract the surveillance video of the surveillance camera.
在上述实施例中，本申请通过事故位置信息查找事故发生地附近的监控摄像，并通过场景图像比对进一步确定实际拍摄到事故发生的监控摄像头，减少人工参与，减轻交警人员的工作压力。In the above embodiment, the present application searches for surveillance cameras near the accident site through accident location information, and further determines the surveillance cameras that actually captured the accident through scene image comparison, thereby reducing manual participation and reducing the work pressure of traffic police officers.
进一步地，所述对所述监控视频进行解析，以确定所述监控视频中的关键帧图像的步骤，具体包括：Further, the step of analyzing the surveillance video to determine key frame images in the surveillance video specifically includes:
按照预设的时间间隔对所述监控视频进行划分，得到若干个视频片段；Divide the surveillance video according to a preset time interval to obtain several video clips;
分别计算每一个所述视频片段中各个视频帧的光流量；Calculate the optical flow of each video frame in each of the video segments respectively;
比对每一个所述视频片段中各个视频帧的光流量，并将光流量最小值对应的视频帧作为所述监控视频的关键帧图像。Compare the optical flow of each video frame in each of the video clips, and use the video frame corresponding to the minimum optical flow as the key frame image of the monitoring video.
其中，基于光流法是一种基于物体运动特征的属性的关键帧提取算法，它一般的实现过程是：在视频镜头中分析物体运动的光流量，每次选择视频镜头中光流移动次数最少的视频帧作为提取到的关键帧。这种方法可以从大部分视频镜头中提取适量的关键帧，提取到的关键帧也可以有效地表达出视频运动的特征。Among them, the optical flow-based method is a key frame extraction algorithm based on the attributes of object motion characteristics. Its general implementation process is: analyze the optical flow of object motion in the video shot, and select the minimum number of optical flow movements in the video shot each time. The video frame is used as the extracted key frame. This method can extract an appropriate amount of key frames from most video shots, and the extracted key frames can also effectively express the characteristics of video motion.
具体的，按照预设的时间间隔(例如5s等)对监控视频进行划分，得到多个视频片段，计算每一个视频片段中各个视频帧的光流量，比对每一个视频片段中各个视频帧的光流量，并将每一个视频片段中光流量最小值对应的视频帧作为该视频片段的关键帧图像，整合所有视频片段的关键帧图像，得到监控视频的关键帧图像。Specifically, the surveillance video is divided according to a preset time interval (for example, 5s, etc.) to obtain multiple video clips, the optical flow of each video frame in each video clip is calculated, and the optical flow of each video frame in each video clip is compared. Optical flow, and take the video frame corresponding to the minimum optical flow in each video clip as the key frame image of the video clip, and integrate the key frame images of all video clips to obtain the key frame image of the monitoring video.
进一步地，利用光流法计算视频帧的光流量公式如下所示：Further, the optical flow formula for calculating the video frame using the optical flow method is as follows:
M(k)＝∑|Lx(i,j,k)|+|Ly(i,j,k)|M(k)=∑|Lx (i,j,k)|+|Ly (i,j,k)|
式中，M(K)表示第k帧的光流量，Lx(i,j,k)表示第k帧像素点(i,j)处光流X的分量，Ly(i,j,k)表示第k帧像素点(i,j)处光流y的分量，计算完成后，取局部最小值作为所要提取的关键帧。In the formula, M(K) represents the optical flow of the kth frame, Lx(i,j,k) represents the component of the optical flow X at the pixel point (i,j) of the kth frame, and Ly(i,j,k) represents The component of the optical flow y at the pixel point (i, j) of the kth frame, after the calculation is completed, take the local minimum value as the key frame to be extracted.
进一步地，所述对所述监控视频进行解析，以确定所述监控视频中的关键帧图像的步骤，具体包括：Further, the step of analyzing the surveillance video to determine key frame images in the surveillance video specifically includes:
计算所述监控视频中相邻两个视频帧的直方图数据和灰度图数据；Calculate the histogram data and grayscale data of two adjacent video frames in the surveillance video;
基于所述直方图数据和所述灰度图数据，计算相邻两个视频帧之间的加权欧式距离；Calculate the weighted Euclidean distance between two adjacent video frames based on the histogram data and the grayscale image data;
基于所述加权欧式距离确定所述监控视频的镜头转换边界；determining the shot transition boundary of the surveillance video based on the weighted Euclidean distance;
基于所述镜头转换边界确定所述监控视频中的关键帧图像。A key frame image in the surveillance video is determined based on the shot transition boundary.
具体的，在本申请另一种实施例中，可以通过镜头转换边界确定监控视频中的关键帧图像。服务器先获取监控视频中相邻两个视频帧的参数，并计算相邻两个视频帧的直方图数据和灰度图数据，其中，直方图数据包括直方图的帧差异值，灰度图数据包括灰度图的均值差值和灰度图的方差差值。对直方图的帧差异值、灰度图的均值差值和灰度图的方差差值进行加权求和，得到相邻两个视频帧的加权欧式距离，通过比对加权欧式距离与预设镜头变化阈值来确定行为记录视频的镜头转换边界，然后根据镜头转换边界确定行为记录视频中的关键帧图像。Specifically, in another embodiment of the present application, the key frame image in the surveillance video may be determined by the shot transition boundary. The server first obtains the parameters of two adjacent video frames in the surveillance video, and calculates the histogram data and grayscale data of the two adjacent video frames, wherein the histogram data includes the frame difference value of the histogram, the grayscale data Including the mean difference of the grayscale image and the variance difference of the grayscale image. The frame difference value of the histogram, the mean difference value of the grayscale image, and the variance difference value of the grayscale image are weighted and summed to obtain the weighted Euclidean distance of two adjacent video frames. By comparing the weighted Euclidean distance with the preset lens Change the threshold to determine the shot transition boundaries of the action recording video, and then determine the key frame images in the action recording video according to the shot transition boundary.
在上述实施例中，本申请分别通过计算视频帧光流量或视频帧镜头转换边界确定监控视频中的关键帧图像，便于后续通过关键帧图像识别事故车辆。In the above embodiments, the present application determines the key frame images in the surveillance video by calculating the video frame optical flow rate or the video frame lens conversion boundary respectively, so as to facilitate subsequent identification of the accident vehicle through the key frame images.
进一步地，所述行驶轨迹识别模型包括输入层、卷积层和输出层，所述在所述监控视频中对所述事故车辆进行标注，并将标注后的监控视频导入预先训练好的行驶轨迹识别模型，得到所述事故车辆的行驶轨迹的步骤，具体包括：Further, the driving trajectory recognition model includes an input layer, a convolution layer and an output layer, the accident vehicle is marked in the monitoring video, and the marked monitoring video is imported into the pre-trained driving trajectory. The steps of identifying the model to obtain the driving trajectory of the accident vehicle specifically include:
通过所述行驶轨迹识别模型的输入层对所述标注后的监控视频进行特征提取，并对提取到的视频特征进行特征向量转化，得到初始特征向量；Feature extraction is performed on the labeled surveillance video through the input layer of the driving track recognition model, and feature vector transformation is performed on the extracted video features to obtain an initial feature vector;
通过所述行驶轨迹识别模型的卷积层对所述初始特征向量进行卷积操作，得到局部特征权值矩阵；Convolution operation is performed on the initial feature vector through the convolution layer of the driving trajectory recognition model to obtain a local feature weight matrix;
通过所述行驶轨迹识别模型的输出层对所述局部特征权值矩阵进行组合，得到所述事故车辆的行驶轨迹。The local feature weight matrix is combined through the output layer of the driving trajectory recognition model to obtain the driving trajectory of the accident vehicle.
具体的，迹识别模型包括行驶轨迹识别模型包括输入层、卷积层和输出层，通过行驶轨迹识别模型的输入层对标注后的监控视频进行特征提取，并对提取到的视频特征进行特征向量转化，得到初始特征向量，通过行驶轨迹识别模型的卷积层对初始特征向量进行卷积操作，得到局部特征权值矩阵，通过行驶轨迹识别模型的的输出层对局部特征权值矩阵进行组合，得到事故车辆的行驶轨迹。Specifically, the track recognition model includes a driving trajectory recognition model including an input layer, a convolution layer and an output layer, and features extraction is performed on the labeled surveillance video through the input layer of the driving trajectory recognition model, and feature vector extraction is performed on the extracted video features. Transform to obtain the initial feature vector, perform the convolution operation on the initial feature vector through the convolution layer of the driving track recognition model, and obtain the local feature weight matrix, and combine the local feature weight matrix through the output layer of the driving track recognition model, Get the trajectory of the accident vehicle.
需要说明的是，卷积计算处理过程为，对于一m*n的矩阵，以1维卷积为例，构建一x*n的卷积核，该卷积核在初始特征向量上滑动运算，得到多个局部特征权值矩阵，初始特征向量是一个多维向量。例如m的值为5，x的值为1，则卷积核自上而下滑动，x首先与第一行的n维向量相乘并求和，得到一个值，随后x继续往下滑动与第2行，第3行…进行卷积运算，共得到5*1的矩阵，即为卷积结果。通过卷积层选取的是监控视频中行驶轨迹的局部特征，输出层把行驶轨迹的局部特征重新通过权值矩阵组装成完整特征，得到事故车辆的行驶轨迹。It should be noted that the convolution calculation process is: for an m*n matrix, taking 1-dimensional convolution as an example, a convolution kernel of x*n is constructed, and the convolution kernel slides on the initial feature vector, Obtain multiple local eigenweight matrices, and the initial eigenvector is a multidimensional vector. For example, the value of m is 5 and the value of x is 1, the convolution kernel slides from top to bottom, x is first multiplied by the n-dimensional vector of the first row and summed to obtain a value, and then x continues to slide down and Line 2, line 3... Perform the convolution operation, and a total of 5*1 matrices are obtained, which is the convolution result. The convolution layer selects the local features of the driving trajectory in the surveillance video, and the output layer assembles the local features of the driving trajectory into a complete feature through the weight matrix to obtain the driving trajectory of the accident vehicle.
在上述实施例中，本申请通过CNN卷积网络模型结构训练一个行驶轨迹识别模型，通过行驶轨迹识别模型实时识别监控视频中事故车辆的车辆轨迹。In the above embodiment, the present application trains a driving trajectory recognition model through the CNN convolutional network model structure, and uses the driving trajectory recognition model to identify the vehicle trajectory of the accident vehicle in the surveillance video in real time.
进一步地，在所述行驶轨迹识别模型包括输入层、卷积层和输出层，所述在所述监控视频中对所述事故车辆进行标注，并将标注后的监控视频导入预先训练好的行驶轨迹识别模型，得到所述事故车辆的行驶轨迹的步骤之前，还包括：Further, the driving trajectory recognition model includes an input layer, a convolution layer and an output layer, the accident vehicle is marked in the monitoring video, and the marked monitoring video is imported into the pre-trained driving. The trajectory recognition model, before the step of obtaining the driving trajectory of the accident vehicle, further includes:
获取训练视频，并对所述训练视频中的行驶车辆进行轨迹标注，得到所述行驶车辆的轨迹标注结果；Acquiring a training video, and labeling the trajectory of the driving vehicle in the training video to obtain a trajectory labeling result of the driving vehicle;
将所述训练视频导入预设的初始识别模型，其中，所述初始识别模型包括输入层、卷积层和输出层；importing the training video into a preset initial recognition model, wherein the initial recognition model includes an input layer, a convolution layer and an output layer;
通过所述初始识别模型的输入层对所述训练视频进行特征提取，并对提取到的视频特征进行特征向量转化，得到训练特征向量；Feature extraction is performed on the training video through the input layer of the initial recognition model, and feature vector transformation is performed on the extracted video features to obtain a training feature vector;
通过所述初始识别模型的卷积层对所述训练特征向量进行卷积操作，得到训练特征权值矩阵；Perform a convolution operation on the training feature vector through the convolution layer of the initial recognition model to obtain a training feature weight matrix;
通过所述初始识别模型的的输出层对所述训练特征权值矩阵进行组合，得到所述行驶车辆的行驶轨迹识别结果；The training feature weight matrix is combined through the output layer of the initial recognition model to obtain the driving trajectory recognition result of the driving vehicle;
基于所述行驶轨迹识别结果与所述轨迹标注结果，使用反向传播算法进行拟合，获取识别误差；Based on the driving trajectory recognition result and the trajectory labeling result, a back-propagation algorithm is used to perform fitting to obtain the recognition error;
将所述识别误差与预设误差阈值进行比较，若所述识别误差大于预设误差阈值，则对所述初始识别模型进行迭代更新，直到所述识别误差小于等于预设误差阈值为止，得到所述行驶轨迹识别模型。The identification error is compared with a preset error threshold, and if the identification error is greater than the preset error threshold, the initial identification model is iteratively updated until the identification error is less than or equal to the preset error threshold, and the result is obtained. The driving trajectory recognition model described above.
其中，反向传播算法，即误差反向传播算法(Backpropagation algorithm，BP算法)适合于多层神经元网络的一种学习算法，它建立在梯度下降法的基础上，用于深度学习网络的误差计算。BP网络的输入、输出关系实质上是一种映射关系：一个n输入m输出的BP神经网络所完成的功能是从n维欧氏空间向m维欧氏空间中一有限域的连续映射，这一映射具有高度非线性。BP算法的学习过程由正向传播过程和反向传播过程组成。在正向传播过程中，输入信息通过输入层经隐含层，逐层处理并传向输出层，并转入反向传播，逐层求出目标函数对各神经元权值的偏导数，构成目标函数对权值向量的梯量，以作为修改权值的依据。Among them, the backpropagation algorithm, that is, the error backpropagation algorithm (Backpropagation algorithm, BP algorithm) is a learning algorithm suitable for multi-layer neuron networks. It is based on the gradient descent method and is used for the error of deep learning networks. calculate. The input and output relationship of BP network is essentially a mapping relationship: the function completed by a BP neural network with n input and m output is a continuous mapping from n-dimensional Euclidean space to a finite field in m-dimensional Euclidean space. A map is highly nonlinear. The learning process of BP algorithm consists of forward propagation process and back propagation process. In the process of forward propagation, the input information is processed layer by layer through the hidden layer through the input layer and transmitted to the output layer, and then transferred to the back propagation, and the partial derivative of the objective function to the weight of each neuron is obtained layer by layer, which constitutes The gradient of the objective function to the weight vector is used as the basis for modifying the weight.
具体的，服务器获取训练视频，训练视频预先收集的车辆行驶视频，并对训练视频中的行驶车辆进行轨迹标注，得到行驶车辆的轨迹标注结果，将训练视频导入预设的初始识别模型，其中，初始识别模型包括输入层、卷积层和输出层，通过初始识别模型的输入层对训练视频进行特征提取，并对提取到的视频特征进行特征向量转化，得到训练特征向量，通过初始识别模型的卷积层对训练特征向量进行卷积操作，得到训练特征权值矩阵，通过初始识别模型的的输出层对训练特征权值矩阵进行组合，得到行驶车辆的行驶轨迹识别结果。Specifically, the server obtains the training video, the vehicle driving video collected in advance by the training video, and labels the trajectory of the driving vehicle in the training video, obtains the trajectory labeling result of the driving vehicle, and imports the training video into the preset initial recognition model, wherein, The initial recognition model includes an input layer, a convolution layer and an output layer. The input layer of the initial recognition model extracts features from the training video, and transforms the extracted video features into feature vectors to obtain training feature vectors. The convolution layer performs the convolution operation on the training feature vector to obtain the training feature weight matrix. The training feature weight matrix is combined through the output layer of the initial recognition model to obtain the driving trajectory recognition result of the driving vehicle.
得到行驶车辆的行驶轨迹识别结果后，基于行驶轨迹识别结果与轨迹标注结果，使用反向传播算法进行拟合，获取识别误差，将识别误差传递到初始识别模型的各个网络层，并将识别误差与预设误差阈值进行比较，若识别误差大于预设误差阈值，则对初始识别模型进行迭代更新，直到识别误差小于等于预设误差阈值为止，得到行驶轨迹识别模型。After obtaining the driving trajectory recognition results of the driving vehicle, based on the driving trajectory recognition results and trajectory labeling results, use the back-propagation algorithm to perform fitting to obtain the recognition error, and transfer the recognition error to each network layer of the initial recognition model, and the recognition error Compared with the preset error threshold, if the recognition error is greater than the preset error threshold, the initial recognition model is iteratively updated until the recognition error is less than or equal to the preset error threshold, and the driving trajectory recognition model is obtained.
进一步地，所述基于所述行驶轨迹、所述行车信息和所述交管法规知识图谱判定所述事故车辆的事故责任的步骤，具体包括：Further, the step of judging the accident responsibility of the accident vehicle based on the driving track, the driving information and the knowledge map of traffic regulations specifically includes:
基于所述行驶轨迹和所述行车信息构建所述事故车辆的关联实体；constructing an associated entity of the accident vehicle based on the driving trajectory and the driving information;
将所述关联实体导入所述交管法规知识图谱，输出所述事故车辆的事故责任判定结果。The associated entity is imported into the traffic regulation knowledge map, and the accident responsibility determination result of the accident vehicle is output.
具体的，服务器基于事故车辆的行驶轨迹和事故车辆的行车信息，构建事故车辆的关联实体，将事故车辆的关联实体导入交管法规知识图谱，输出事故车辆的事故责任判定结果。在本申请一种具体的实施例中，X路段上的某次事故构建的关联实体如下：“甲车—速度96”、“乙车—56”和“X路段-限速80”，将上述关联实体导入交管法规知识图谱，输出结果为“甲车—超速违规”，“乙车—行驶正常”，根据输出结果对本次事故进行定责，确定甲车负事故全责。Specifically, the server constructs the associated entity of the accident vehicle based on the driving track of the accident vehicle and the driving information of the accident vehicle, imports the associated entity of the accident vehicle into the traffic control law knowledge map, and outputs the accident responsibility determination result of the accident vehicle. In a specific embodiment of the present application, the associated entities constructed for an accident on section X are as follows: "Car A - Speed 96", "Car B - 56", and "Section X - Speed Limit 80". The related entity imports the knowledge map of traffic control regulations, and the output results are "Vehicle A - speeding violation" and "Vehicle B - normal driving". According to the output results, the responsibility for the accident is determined, and it is determined that car A is fully responsible for the accident.
在本实施例中，本申请公开了一种交通事故定责的方法，属于人工智能技术领域。本申请通过交通事故的位置自动查找交通事故对应的监控视频，减少交警人员的工作量，然后对监控视频进行解析，以确定监控视频中的关键帧图像，对关键帧图像进行内容识别，确定交通事故中的事故车辆，然后在监控视频中对标注事故车辆，并将标注后的监控视频导入预先训练好的行驶轨迹识别模型，得到事故车辆的行驶轨迹，最后通过事故车辆的行驶轨迹、行车信息和预先构建的交管法规知识图谱自动判定事故车辆的事故责任。本申请通过事故定位和监控视频信息识别获取事故位置信息和事故车辆信息，并结合交管法规知识图谱实现对交通事故的自动定责，减轻交警人员的工作压力，提高交通事故案件处理效率。In this embodiment, the present application discloses a method for determining responsibility for a traffic accident, which belongs to the technical field of artificial intelligence. This application automatically finds the surveillance video corresponding to the traffic accident through the location of the traffic accident, reduces the workload of the traffic police, and then parses the surveillance video to determine the key frame images in the surveillance video, performs content recognition on the key frame images, and determines the traffic The accident vehicle in the accident, then mark the accident vehicle in the surveillance video, and import the labeled surveillance video into the pre-trained driving trajectory recognition model to obtain the driving trajectory of the accident vehicle, and finally pass the driving trajectory and driving information of the accident vehicle. And the pre-built traffic control law knowledge map automatically determines the accident responsibility of the accident vehicle. The application obtains accident location information and accident vehicle information through accident location and monitoring video information identification, and combines the knowledge map of traffic control laws and regulations to realize automatic responsibility determination for traffic accidents, reduce the work pressure of traffic police personnel, and improve the efficiency of traffic accident case handling.
需要强调的是，为进一步保证上述监控视频的私密和安全性，上述监控视频还可以存储于一区块链的节点中。It should be emphasized that, in order to further ensure the privacy and security of the above-mentioned surveillance video, the above-mentioned surveillance video can also be stored in a node of a blockchain.
本申请所指区块链是分布式数据存储、点对点传输、共识机制、加密算法等计算机技术的新型应用模式。区块链(Blockchain)，本质上是一个去中心化的数据库，是一串使用密码学方法相关联产生的数据块，每一个数据块中包含了一批次网络交易的信息，用于验证其信息的有效性(防伪)和生成下一个区块。区块链可以包括区块链底层平台、平台产品服务层以及应用服务层等。The blockchain referred to in this application is a new application mode of computer technologies such as distributed data storage, point-to-point transmission, consensus mechanism, and encryption algorithm. Blockchain, essentially a decentralized database, is a series of data blocks associated with cryptographic methods. Each data block contains a batch of network transaction information to verify its Validity of information (anti-counterfeiting) and generation of the next block. The blockchain can include the underlying platform of the blockchain, the platform product service layer, and the application service layer.
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分流程，是可以通过计算机可读指令来指令相关的硬件来完成，该计算机可读指令可存储于一计算机可读取存储介质中，该计算机可读指令在执行时，可包括如上述各方法的实施例的流程。其中，前述的存储介质可为磁碟、光盘、只读存储记忆体(Read-Only Memory，ROM)等非易失性存储介质，或随机存储记忆体(Random Access Memory，RAM)等。Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be implemented by instructing relevant hardware through computer-readable instructions, and the computer-readable instructions can be stored in a computer-readable storage medium. , when the computer-readable instructions are executed, the processes of the above-mentioned method embodiments may be included. The aforementioned storage medium may be a non-volatile storage medium such as a magnetic disk, an optical disk, a read-only memory (Read-Only Memory, ROM), or a random access memory (Random Access Memory, RAM).
应该理解的是，虽然附图的流程图中的各个步骤按照箭头的指示依次显示，但是这些步骤并不是必然按照箭头指示的顺序依次执行。除非本文中有明确的说明，这些步骤的执行并没有严格的顺序限制，其可以以其他的顺序执行。而且，附图的流程图中的至少一部分步骤可以包括多个子步骤或者多个阶段，这些子步骤或者阶段并不必然是在同一时刻执行完成，而是可以在不同的时刻执行，其执行顺序也不必然是依次进行，而是可以与其他步骤或者其他步骤的子步骤或者阶段的至少一部分轮流或者交替地执行。It should be understood that although the various steps in the flowchart of the accompanying drawings are sequentially shown in the order indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless explicitly stated herein, the execution of these steps is not strictly limited to the order and may be performed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but may be executed at different times, and the execution sequence is also It does not have to be performed sequentially, but may be performed alternately or alternately with other steps or at least a portion of sub-steps or stages of other steps.
进一步参考图3，作为对上述图2所示方法的实现，本申请提供了一种交通事故定责装置的一个实施例，该装置实施例与图2所示的方法实施例相对应，该装置具体可以应用于各种电子设备中。Further referring to FIG. 3 , as an implementation of the method shown in FIG. 2 above, the present application provides an embodiment of a device for determining responsibility for a traffic accident. The device embodiment corresponds to the method embodiment shown in FIG. 2 . Specifically, it can be applied to various electronic devices.
如图3所示，本实施例所述的交通事故定责装置包括：As shown in FIG. 3 , the device for determining responsibility for a traffic accident in this embodiment includes:
事故定位模块301，用于接收事故定责指令，根据所述事故定责指令对事故的位置进行定位，得到事故位置信息；The
信息查询模块302，用于获取事故发生时间，根据所述事故位置信息和所述事故发生时间查找所述事故对应的监控视频；An
视频解析模块303，用于对所述监控视频进行解析，以确定所述监控视频中的关键帧图像；A
内容识别模块304，用于对所述关键帧图像进行内容识别，确定所述关键帧图像中的事故车辆，并获取所述事故车辆的行车信息；A
轨迹识别模块305，用于在所述监控视频中对所述事故车辆进行标注，并将标注后的监控视频导入预先训练好的行驶轨迹识别模型，得到所述事故车辆的行驶轨迹；The
事故定责模块306，用于获取预先构建的交管法规知识图谱，并基于所述行驶轨迹、所述行车信息和所述交管法规知识图谱判定所述事故车辆的事故责任。The
进一步地，所述信息查询模块302具体包括：Further, the
摄像头获取单元，用于根据所述事故位置信息查找事故发生地附近预设范围内的监控摄像头；a camera acquisition unit, configured to search for surveillance cameras within a preset range near the accident site according to the accident location information;
视频查询单元，用于调用所述监控摄像头的监控影像，并在所述监控影像中截取所述事故发生时间前后预设时长范围内的影像片段，得到所述事故对应的监控视频。The video query unit is used for calling the monitoring image of the monitoring camera, and intercepting the video clips within the preset time range before and after the occurrence time of the accident from the monitoring image, so as to obtain the monitoring video corresponding to the accident.
进一步地，所述视频解析模块303具体包括：Further, the
视频分割单元，用于按照预设的时间间隔对所述监控视频进行划分，得到若干个视频片段；a video segmentation unit, configured to divide the surveillance video according to a preset time interval to obtain several video clips;
光流量计算单元，用于分别计算每一个所述视频片段中各个视频帧的光流量；an optical flow calculation unit, configured to calculate the optical flow of each video frame in each of the video segments respectively;
第一关键帧获取单元，用于比对每一个所述视频片段中各个视频帧的光流量，并将光流量最小值对应的视频帧作为所述监控视频的关键帧图像。The first key frame acquisition unit is configured to compare the optical flow of each video frame in each of the video clips, and use the video frame corresponding to the minimum optical flow as the key frame image of the monitoring video.
进一步地，所述视频解析模块303具体包括：Further, the
视频数据获取单元，用于计算所述监控视频中相邻两个视频帧的直方图数据和灰度图数据；A video data acquisition unit, used to calculate the histogram data and grayscale data of two adjacent video frames in the surveillance video;
欧式距离计算单元，用于基于所述直方图数据和所述灰度图数据，计算相邻两个视频帧之间的加权欧式距离；an Euclidean distance calculation unit for calculating the weighted Euclidean distance between two adjacent video frames based on the histogram data and the grayscale image data;
镜头转换边界单元，用于基于所述加权欧式距离确定所述监控视频的镜头转换边界；a shot transition boundary unit for determining a shot transition boundary of the surveillance video based on the weighted Euclidean distance;
第二关键帧获取单元，用于基于所述镜头转换边界确定所述监控视频中的关键帧图像。A second key frame acquiring unit, configured to determine a key frame image in the surveillance video based on the shot transition boundary.
进一步地，所述行驶轨迹识别模型包括输入层、卷积层和输出层，所述轨迹识别模块305具体包括：Further, the driving trajectory recognition model includes an input layer, a convolution layer and an output layer, and the
特征预处理单元，用于通过所述行驶轨迹识别模型的输入层对所述标注后的监控视频进行特征提取，并对提取到的视频特征进行特征向量转化，得到初始特征向量；a feature preprocessing unit, configured to perform feature extraction on the labeled surveillance video through the input layer of the driving track recognition model, and perform feature vector transformation on the extracted video features to obtain an initial feature vector;
卷积单元，用于通过所述行驶轨迹识别模型的卷积层对所述初始特征向量进行卷积操作，得到局部特征权值矩阵；a convolution unit, configured to perform a convolution operation on the initial feature vector through the convolution layer of the driving trajectory recognition model to obtain a local feature weight matrix;
全连接单元，用于通过所述行驶轨迹识别模型的输出层对所述局部特征权值矩阵进行组合，得到所述事故车辆的行驶轨迹。The fully connected unit is used for combining the local feature weight matrix through the output layer of the driving trajectory recognition model to obtain the driving trajectory of the accident vehicle.
进一步地，在所述行驶轨迹识别模型包括输入层、卷积层和输出层，所述交通事故定责的装置还包括：Further, the driving trajectory recognition model includes an input layer, a convolution layer and an output layer, and the device for determining responsibility for a traffic accident further includes:
训练数据获取模块，用于获取训练视频，并对所述训练视频中的行驶车辆进行轨迹标注，得到所述行驶车辆的轨迹标注结果；A training data acquisition module, configured to acquire a training video, and label the trajectory of the driving vehicle in the training video to obtain a trajectory annotation result of the driving vehicle;
训练数据导入模块，用于将所述训练视频导入预设的初始识别模型，其中，所述初始识别模型包括输入层、卷积层和输出层；A training data importing module for importing the training video into a preset initial recognition model, wherein the initial recognition model includes an input layer, a convolution layer and an output layer;
第一训练模块，用于通过所述初始识别模型的输入层对所述训练视频进行特征提取，并对提取到的视频特征进行特征向量转化，得到训练特征向量；The first training module is used to perform feature extraction on the training video through the input layer of the initial recognition model, and perform feature vector transformation on the extracted video features to obtain a training feature vector;
第二训练模块，用于通过所述初始识别模型的卷积层对所述训练特征向量进行卷积操作，得到训练特征权值矩阵；The second training module is used to perform a convolution operation on the training feature vector through the convolution layer of the initial recognition model to obtain a training feature weight matrix;
第三训练模块，用于通过所述初始识别模型的的输出层对所述训练特征权值矩阵进行组合，得到所述行驶车辆的行驶轨迹识别结果；a third training module, configured to combine the training feature weight matrix through the output layer of the initial recognition model to obtain the driving trajectory recognition result of the driving vehicle;
反向拟合模块，用于基于所述行驶轨迹识别结果与所述轨迹标注结果，使用反向传播算法进行拟合，获取识别误差；a reverse fitting module, configured to use a back-propagation algorithm to perform fitting based on the driving trajectory recognition result and the trajectory labeling result to obtain a recognition error;
迭代训练模块，用于将所述识别误差与预设误差阈值进行比较，若所述识别误差大于预设误差阈值，则对所述初始识别模型进行迭代更新，直到所述识别误差小于等于预设误差阈值为止，得到所述行驶轨迹识别模型。an iterative training module, configured to compare the recognition error with a preset error threshold, and if the recognition error is greater than a preset error threshold, iteratively update the initial recognition model until the recognition error is less than or equal to a preset Up to the error threshold, the driving trajectory recognition model is obtained.
进一步地，所述事故定责模块306具体包括：Further, the
实体构建单元，用于基于所述行驶轨迹和所述行车信息构建所述事故车辆的关联实体；an entity construction unit, configured to construct an associated entity of the accident vehicle based on the driving track and the driving information;
事故定责单元，用于将所述关联实体导入所述交管法规知识图谱，输出所述事故车辆的事故责任判定结果。The accident responsibility determination unit is used for importing the associated entity into the traffic control law knowledge map, and outputting the accident responsibility determination result of the accident vehicle.
在本实施例中，本申请公开了一种交通事故定责的装置，属于人工智能技术领域。本申请通过交通事故的位置自动查找交通事故对应的监控视频，减少交警人员的工作量，然后对监控视频进行解析，以确定监控视频中的关键帧图像，对关键帧图像进行内容识别，确定交通事故中的事故车辆，然后在监控视频中对标注事故车辆，并将标注后的监控视频导入预先训练好的行驶轨迹识别模型，得到事故车辆的行驶轨迹，最后通过事故车辆的行驶轨迹、行车信息和预先构建的交管法规知识图谱自动判定事故车辆的事故责任。本申请通过事故定位和监控视频信息识别获取事故位置信息和事故车辆信息，并结合交管法规知识图谱实现对交通事故的自动定责，减轻交警人员的工作压力，提高交通事故案件处理效率。In this embodiment, the present application discloses a device for determining responsibility for a traffic accident, which belongs to the technical field of artificial intelligence. This application automatically finds the surveillance video corresponding to the traffic accident through the location of the traffic accident, reduces the workload of the traffic police, and then parses the surveillance video to determine the key frame images in the surveillance video, performs content recognition on the key frame images, and determines the traffic The accident vehicle in the accident, then mark the accident vehicle in the surveillance video, and import the labeled surveillance video into the pre-trained driving trajectory recognition model to obtain the driving trajectory of the accident vehicle, and finally pass the driving trajectory and driving information of the accident vehicle. And the pre-built traffic control law knowledge map automatically determines the accident responsibility of the accident vehicle. The application obtains accident location information and accident vehicle information through accident location and monitoring video information identification, and combines the knowledge map of traffic control laws and regulations to realize automatic responsibility determination for traffic accidents, reduce the work pressure of traffic police personnel, and improve the efficiency of traffic accident case handling.
为解决上述技术问题，本申请实施例还提供计算机设备。具体请参阅图4，图4为本实施例计算机设备基本结构框图。To solve the above technical problems, the embodiments of the present application further provide computer equipment. Please refer to FIG. 4 for details. FIG. 4 is a block diagram of a basic structure of a computer device according to this embodiment.
所述计算机设备4包括通过系统总线相互通信连接存储器41、处理器42、网络接口43。需要指出的是，图中仅示出了具有组件41-43的计算机设备4，但是应理解的是，并不要求实施所有示出的组件，可以替代的实施更多或者更少的组件。其中，本技术领域技术人员可以理解，这里的计算机设备是一种能够按照事先设定或存储的指令，自动进行数值计算和/或信息处理的设备，其硬件包括但不限于微处理器、专用集成电路(ApplicationSpecific Integrated Circuit，ASIC)、可编程门阵列(Field－Programmable GateArray，FPGA)、数字处理器(Digital Signal Processor，DSP)、嵌入式设备等。The computer device 4 includes a
所述计算机设备可以是桌上型计算机、笔记本、掌上电脑及云端服务器等计算设备。所述计算机设备可以与用户通过键盘、鼠标、遥控器、触摸板或声控设备等方式进行人机交互。The computer equipment may be a desktop computer, a notebook computer, a palmtop computer, a cloud server and other computing equipment. The computer device can perform human-computer interaction with the user through a keyboard, a mouse, a remote control, a touch pad or a voice control device.
所述存储器41至少包括一种类型的可读存储介质，所述可读存储介质包括闪存、硬盘、多媒体卡、卡型存储器(例如，SD或DX存储器等)、随机访问存储器(RAM)、静态随机访问存储器(SRAM)、只读存储器(ROM)、电可擦除可编程只读存储器(EEPROM)、可编程只读存储器(PROM)、磁性存储器、磁盘、光盘等。在一些实施例中，所述存储器41可以是所述计算机设备4的内部存储单元，例如该计算机设备4的硬盘或内存。在另一些实施例中，所述存储器41也可以是所述计算机设备4的外部存储设备，例如该计算机设备4上配备的插接式硬盘，智能存储卡(Smart Media Card,SMC)，安全数字(Secure Digital,SD)卡，闪存卡(FlashCard)等。当然，所述存储器41还可以既包括所述计算机设备4的内部存储单元也包括其外部存储设备。本实施例中，所述存储器41通常用于存储安装于所述计算机设备4的操作系统和各类应用软件，例如交通事故定责方法的计算机可读指令等。此外，所述存储器41还可以用于暂时地存储已经输出或者将要输出的各类数据。The
所述处理器42在一些实施例中可以是中央处理器(Central Processing Unit，CPU)、控制器、微控制器、微处理器、或其他数据处理芯片。该处理器42通常用于控制所述计算机设备4的总体操作。本实施例中，所述处理器42用于运行所述存储器41中存储的计算机可读指令或者处理数据，例如运行所述交通事故定责方法的计算机可读指令。The
所述网络接口43可包括无线网络接口或有线网络接口，该网络接口43通常用于在所述计算机设备4与其他电子设备之间建立通信连接。The
本申请公开了一种计算机设备，属于人工智能技术领域。本申请通过交通事故的位置自动查找交通事故对应的监控视频，减少交警人员的工作量，然后对监控视频进行解析，以确定监控视频中的关键帧图像，对关键帧图像进行内容识别，确定交通事故中的事故车辆，然后在监控视频中对标注事故车辆，并将标注后的监控视频导入预先训练好的行驶轨迹识别模型，得到事故车辆的行驶轨迹，最后通过事故车辆的行驶轨迹、行车信息和预先构建的交管法规知识图谱自动判定事故车辆的事故责任。本申请通过事故定位和监控视频信息识别获取事故位置信息和事故车辆信息，并结合交管法规知识图谱实现对交通事故的自动定责，减轻交警人员的工作压力，提高交通事故案件处理效率。The application discloses a computer device, which belongs to the technical field of artificial intelligence. This application automatically finds the surveillance video corresponding to the traffic accident through the location of the traffic accident, reduces the workload of the traffic police, and then parses the surveillance video to determine the key frame images in the surveillance video, performs content recognition on the key frame images, and determines the traffic The accident vehicle in the accident, then mark the accident vehicle in the surveillance video, and import the labeled surveillance video into the pre-trained driving trajectory recognition model to obtain the driving trajectory of the accident vehicle, and finally pass the driving trajectory and driving information of the accident vehicle. And the pre-built traffic control law knowledge map automatically determines the accident responsibility of the accident vehicle. The application obtains accident location information and accident vehicle information through accident location and monitoring video information identification, and combines the knowledge map of traffic control laws and regulations to realize automatic responsibility determination for traffic accidents, reduce the work pressure of traffic police personnel, and improve the efficiency of traffic accident case handling.
本申请还提供了另一种实施方式，即提供一种计算机可读存储介质，所述计算机可读存储介质存储有计算机可读指令，所述计算机可读指令可被至少一个处理器执行，以使所述至少一个处理器执行如上述的交通事故定责方法的步骤。The present application also provides another embodiment, that is, to provide a computer-readable storage medium, where the computer-readable storage medium stores computer-readable instructions, and the computer-readable instructions can be executed by at least one processor to The at least one processor is caused to execute the steps of the traffic accident liability method as described above.
本申请公开了一种存储介质，属于人工智能技术领域。本申请通过交通事故的位置自动查找交通事故对应的监控视频，减少交警人员的工作量，然后对监控视频进行解析，以确定监控视频中的关键帧图像，对关键帧图像进行内容识别，确定交通事故中的事故车辆，然后在监控视频中对标注事故车辆，并将标注后的监控视频导入预先训练好的行驶轨迹识别模型，得到事故车辆的行驶轨迹，最后通过事故车辆的行驶轨迹、行车信息和预先构建的交管法规知识图谱自动判定事故车辆的事故责任。本申请通过事故定位和监控视频信息识别获取事故位置信息和事故车辆信息，并结合交管法规知识图谱实现对交通事故的自动定责，减轻交警人员的工作压力，提高交通事故案件处理效率。The application discloses a storage medium, which belongs to the technical field of artificial intelligence. This application automatically finds the surveillance video corresponding to the traffic accident through the location of the traffic accident, reduces the workload of the traffic police, and then parses the surveillance video to determine the key frame images in the surveillance video, performs content recognition on the key frame images, and determines the traffic The accident vehicle in the accident, then mark the accident vehicle in the surveillance video, and import the labeled surveillance video into the pre-trained driving trajectory recognition model to obtain the driving trajectory of the accident vehicle, and finally pass the driving trajectory and driving information of the accident vehicle. And the pre-built traffic control law knowledge map automatically determines the accident responsibility of the accident vehicle. The application obtains accident location information and accident vehicle information through accident location and monitoring video information identification, and combines the knowledge map of traffic control laws and regulations to realize automatic responsibility determination for traffic accidents, reduce the work pressure of traffic police personnel, and improve the efficiency of traffic accident case handling.
通过以上的实施方式的描述，本领域的技术人员可以清楚地了解到上述实施例方法可借助软件加必需的通用硬件平台的方式来实现，当然也可以通过硬件，但很多情况下前者是更佳的实施方式。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质(如ROM/RAM、磁碟、光盘)中，包括若干指令用以使得一台终端设备(可以是手机，计算机，服务器，空调器，或者网络设备等)执行本申请各个实施例所述的方法。From the description of the above embodiments, those skilled in the art can clearly understand that the method of the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is better implementation. Based on this understanding, the technical solution of the present application can be embodied in the form of a software product in essence or in a part that contributes to the prior art, and the computer software product is stored in a storage medium (such as ROM/RAM, magnetic disk, CD-ROM), including several instructions to make a terminal device (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) execute the methods described in the various embodiments of this application.
本申请可用于众多通用或专用的计算机系统环境或配置中。例如：个人计算机、服务器计算机、手持设备或便携式设备、平板型设备、多处理器系统、基于微处理器的系统、置顶盒、可编程的消费电子设备、网络PC、小型计算机、大型计算机、包括以上任何系统或设备的分布式计算环境等等。本申请可以在由计算机执行的计算机可执行指令的一般上下文中描述，例如程序模块。一般地，程序模块包括执行特定任务或实现特定抽象数据类型的例程、程序、对象、组件、数据结构等等。也可以在分布式计算环境中实践本申请，在这些分布式计算环境中，由通过通信网络而被连接的远程处理设备来执行任务。在分布式计算环境中，程序模块可以位于包括存储设备在内的本地和远程计算机存储介质中。The present application may be used in numerous general purpose or special purpose computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, including A distributed computing environment for any of the above systems or devices, and the like. The application may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The application may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including storage devices.
显然，以上所描述的实施例仅仅是本申请一部分实施例，而不是全部的实施例，附图中给出了本申请的较佳实施例，但并不限制本申请的专利范围。本申请可以以许多不同的形式来实现，相反地，提供这些实施例的目的是使对本申请的公开内容的理解更加透彻全面。尽管参照前述实施例对本申请进行了详细的说明，对于本领域的技术人员来而言，其依然可以对前述各具体实施方式所记载的技术方案进行修改，或者对其中部分技术特征进行等效替换。凡是利用本申请说明书及附图内容所做的等效结构，直接或间接运用在其他相关的技术领域，均同理在本申请专利保护范围之内。Obviously, the above-described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. The accompanying drawings show the preferred embodiments of the present application, but do not limit the scope of the patent of the present application. This application may be embodied in many different forms, rather these embodiments are provided so that a thorough and complete understanding of the disclosure of this application is provided. Although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing specific embodiments, or perform equivalent replacements for some of the technical features. . Any equivalent structure made by using the contents of the description and drawings of the present application, which is directly or indirectly used in other related technical fields, is also within the scope of protection of the patent of the present application.
Claims (10)
Translated fromChinesePriority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210171452.0ACN114550053A (en) | 2022-02-24 | 2022-02-24 | A method, device, computer equipment and storage medium for determining responsibility for a traffic accident |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210171452.0ACN114550053A (en) | 2022-02-24 | 2022-02-24 | A method, device, computer equipment and storage medium for determining responsibility for a traffic accident |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114550053Atrue CN114550053A (en) | 2022-05-27 |
Family
ID=81677338
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210171452.0APendingCN114550053A (en) | 2022-02-24 | 2022-02-24 | A method, device, computer equipment and storage medium for determining responsibility for a traffic accident |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114550053A (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115115822A (en)* | 2022-06-30 | 2022-09-27 | 小米汽车科技有限公司 | Vehicle-end image processing method and device, vehicle, storage medium and chip |
| CN115424431A (en)* | 2022-07-11 | 2022-12-02 | 广州国交润万交通信息有限公司 | System for realizing equipment monitoring through field personnel positioning |
| CN115690702A (en)* | 2023-01-05 | 2023-02-03 | 南京云创大数据科技股份有限公司 | Road accident reconstruction and assistance responsibility determination method based on image segmentation algorithm |
| CN116403394A (en)* | 2022-12-05 | 2023-07-07 | 合众新能源汽车股份有限公司 | Traffic accident determination method, device and related equipment based on central control system |
| CN116579550A (en)* | 2023-04-27 | 2023-08-11 | 丹阳市公安局 | A 5G visual alarm service method and system |
| CN117808437A (en)* | 2024-02-28 | 2024-04-02 | 山东金宇信息科技集团有限公司 | Traffic management method, equipment and medium based on virtual simulation technology |
| CN117894182A (en)* | 2024-03-15 | 2024-04-16 | 长春师范大学 | A method and system for quickly collecting vehicle accident data based on Internet of Vehicles |
| CN117933948A (en)* | 2024-03-21 | 2024-04-26 | 贵州华鑫信息技术有限公司 | Vehicle accident handling method, device, equipment and medium |
| CN118171781A (en)* | 2024-05-13 | 2024-06-11 | 东南大学 | A method and system for intelligent detection of motor vehicle accidents on highways based on real-time trajectory prediction |
| CN119206593A (en)* | 2024-11-29 | 2024-12-27 | 天翼交通科技有限公司 | A video-based traffic accident analysis method, device, equipment and medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101360184A (en)* | 2008-09-22 | 2009-02-04 | 腾讯科技（深圳）有限公司 | System and method for extracting key frame of video |
| CN108986474A (en)* | 2018-08-01 | 2018-12-11 | 平安科技（深圳）有限公司 | Fix duty method, apparatus, computer equipment and the computer storage medium of traffic accident |
| CN111723672A (en)* | 2020-05-25 | 2020-09-29 | 华南理工大学 | A method, device and storage medium for obtaining a video recognition driving track |
| CN212009589U (en)* | 2020-04-15 | 2020-11-24 | 华南理工大学 | A deep learning-based video recognition driving trajectory acquisition device |
| CN113674523A (en)* | 2020-05-14 | 2021-11-19 | 华为技术有限公司 | Traffic accident analysis method, device and equipment |
| CN113766330A (en)* | 2021-05-26 | 2021-12-07 | 腾讯科技（深圳）有限公司 | Method and device for generating recommendation information based on video |
- 2022
- 2022-02-24CNCN202210171452.0Apatent/CN114550053A/enactivePending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101360184A (en)* | 2008-09-22 | 2009-02-04 | 腾讯科技（深圳）有限公司 | System and method for extracting key frame of video |
| CN108986474A (en)* | 2018-08-01 | 2018-12-11 | 平安科技（深圳）有限公司 | Fix duty method, apparatus, computer equipment and the computer storage medium of traffic accident |
| WO2020024457A1 (en)* | 2018-08-01 | 2020-02-06 | 平安科技（深圳）有限公司 | Liability cognizance method and device of traffic accident and computer readable storage medium |
| CN212009589U (en)* | 2020-04-15 | 2020-11-24 | 华南理工大学 | A deep learning-based video recognition driving trajectory acquisition device |
| CN113674523A (en)* | 2020-05-14 | 2021-11-19 | 华为技术有限公司 | Traffic accident analysis method, device and equipment |
| CN111723672A (en)* | 2020-05-25 | 2020-09-29 | 华南理工大学 | A method, device and storage medium for obtaining a video recognition driving track |
| CN113766330A (en)* | 2021-05-26 | 2021-12-07 | 腾讯科技（深圳）有限公司 | Method and device for generating recommendation information based on video |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115115822A (en)* | 2022-06-30 | 2022-09-27 | 小米汽车科技有限公司 | Vehicle-end image processing method and device, vehicle, storage medium and chip |
| CN115115822B (en)* | 2022-06-30 | 2023-10-31 | 小米汽车科技有限公司 | Vehicle-end image processing method and device, vehicle, storage medium and chip |
| CN115424431A (en)* | 2022-07-11 | 2022-12-02 | 广州国交润万交通信息有限公司 | System for realizing equipment monitoring through field personnel positioning |
| CN116403394A (en)* | 2022-12-05 | 2023-07-07 | 合众新能源汽车股份有限公司 | Traffic accident determination method, device and related equipment based on central control system |
| CN115690702A (en)* | 2023-01-05 | 2023-02-03 | 南京云创大数据科技股份有限公司 | Road accident reconstruction and assistance responsibility determination method based on image segmentation algorithm |
| CN116579550A (en)* | 2023-04-27 | 2023-08-11 | 丹阳市公安局 | A 5G visual alarm service method and system |
| CN117808437A (en)* | 2024-02-28 | 2024-04-02 | 山东金宇信息科技集团有限公司 | Traffic management method, equipment and medium based on virtual simulation technology |
| CN117808437B (en)* | 2024-02-28 | 2024-05-17 | 山东金宇信息科技集团有限公司 | Traffic management method, equipment and medium based on virtual simulation technology |
| CN117894182A (en)* | 2024-03-15 | 2024-04-16 | 长春师范大学 | A method and system for quickly collecting vehicle accident data based on Internet of Vehicles |
| CN117933948A (en)* | 2024-03-21 | 2024-04-26 | 贵州华鑫信息技术有限公司 | Vehicle accident handling method, device, equipment and medium |
| CN118171781A (en)* | 2024-05-13 | 2024-06-11 | 东南大学 | A method and system for intelligent detection of motor vehicle accidents on highways based on real-time trajectory prediction |
| CN118171781B (en)* | 2024-05-13 | 2024-08-13 | 东南大学 | A method and system for intelligent detection of highway motor vehicle accidents based on real-time trajectory prediction |
| CN119206593A (en)* | 2024-11-29 | 2024-12-27 | 天翼交通科技有限公司 | A video-based traffic accident analysis method, device, equipment and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114550053A (en) | A method, device, computer equipment and storage medium for determining responsibility for a traffic accident | |
| Chen et al. | An edge traffic flow detection scheme based on deep learning in an intelligent transportation system | |
| CN111797893B (en) | Neural network training method, image classification system and related equipment | |
| CN112329659B (en) | Weak supervision semantic segmentation method based on vehicle image and related equipment thereof | |
| CN110458107B (en) | Method and device for image recognition | |
| WO2022142014A1 (en) | Multi-modal information fusion-based text classification method, and related device thereof | |
| WO2021139191A1 (en) | Method for data labeling and apparatus for data labeling | |
| CN113704531A (en) | Image processing method, image processing device, electronic equipment and computer readable storage medium | |
| CN113807399A (en) | Neural network training method, neural network detection method and neural network detection device | |
| CN113011568A (en) | Model training method, data processing method and equipment | |
| CN112434746B (en) | Pre-labeling method based on hierarchical migration learning and related equipment thereof | |
| CN113139490B (en) | Image feature matching method and device, computer equipment and storage medium | |
| CN113570512B (en) | Image data processing method, computer and readable storage medium | |
| CN114550051A (en) | Vehicle loss detection method and device, computer equipment and storage medium | |
| CN114550052B (en) | Vehicle accident handling method, device, computer equipment and storage medium | |
| CN114241411B (en) | Counting model processing method and device based on target detection and computer equipment | |
| CN113705293A (en) | Image scene recognition method, device, equipment and readable storage medium | |
| CN113641797A (en) | Data processing method, device, equipment, storage medium and computer program product | |
| CN115129908A (en) | A model optimization method, device, equipment, storage medium and program product | |
| CN117235257A (en) | Emotion prediction method, device, equipment and storage medium based on artificial intelligence | |
| CN115941322A (en) | Attack detection method, device, equipment and storage medium based on artificial intelligence | |
| CN117132950A (en) | A vehicle tracking method, system, equipment and storage medium | |
| WO2024054576A1 (en) | System and method synthetic data generation | |
| CN118552826A (en) | Visible light and infrared image target detection method and device based on dual-stream attention | |
| CN111091198B (en) | Data processing method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |