CN115965657B - Target tracking method, electronic device, storage medium and vehicle - Google Patents
Target tracking method, electronic device, storage medium and vehicleDownload PDFInfo
- Publication number
- CN115965657B CN115965657BCN202310171954.8ACN202310171954ACN115965657BCN 115965657 BCN115965657 BCN 115965657BCN 202310171954 ACN202310171954 ACN 202310171954ACN 115965657 BCN115965657 BCN 115965657B
- Authority
- CN
- China
- Prior art keywords
- frame
- predicted trajectory
- queue
- predicted
- current frame
- 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.)
- Active
Links
Images
Classifications
- 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
- G06N3/0985—Hyperparameter optimisation; Meta-learning; Learning-to-learn
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Biophysics (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Multimedia (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
Translated fromChinese
Description
Translated fromChinese技术领域technical field
本发明涉及计算机技术领域，具体提供一种目标跟踪方法、电子设备、存储介质及车辆。The invention relates to the field of computer technology, and specifically provides a target tracking method, electronic equipment, a storage medium and a vehicle.
背景技术Background technique
目前，高级辅助驾驶的功能越来越受到大家的关注，多目标跟踪是高级辅助驾驶系统中的基础一环。常用的目标跟踪方法大多为光流法、滤波类算法，其中光流法受环境影响较大，在复杂环境下跟踪的准确度较差，滤波类算法对于目标形态变化、遮挡或目标消失等情况下跟踪的准确度较差。At present, the function of advanced assisted driving has attracted more and more attention, and multi-target tracking is a basic part of the advanced assisted driving system. Most of the commonly used target tracking methods are optical flow method and filtering algorithm. Among them, the optical flow method is greatly affected by the environment, and the tracking accuracy in complex environments is poor. Lower tracking accuracy is poor.
相应地，本领域需要一种新的目标跟踪方案来解决上述问题。Correspondingly, a new target tracking solution is needed in the field to solve the above problems.
发明内容Contents of the invention
为了克服上述缺陷，提出了本发明，以提供解决或至少部分地解决上述的技术问题。本发明提供了一种目标跟踪方法、电子设备、存储介质及车辆。In order to overcome the above drawbacks, the present invention is proposed to solve or at least partly solve the above technical problems. The invention provides a target tracking method, electronic equipment, a storage medium and a vehicle.
在第一方面，本发明提供一种目标跟踪方法，所述方法包括：获取车载传感器采集的传感器数据；将所述传感器数据输入网络模型，输出至少一个检测目标的当前帧检测框和当前帧预测轨迹信息；基于所述当前帧检测框、所述当前帧预测轨迹信息和历史帧预测轨迹信息对所述至少一个检测目标进行跟踪，获得跟踪结果。In a first aspect, the present invention provides a target tracking method, the method comprising: acquiring sensor data collected by vehicle sensors; inputting the sensor data into a network model, and outputting at least one current frame detection frame and current frame prediction of a detected target Trajectory information: track the at least one detection target based on the current frame detection frame, the current frame predicted trajectory information, and historical frame predicted trajectory information, and obtain a tracking result.
在一个实施方式中，所述基于所述当前帧检测框、所述当前帧预测轨迹信息和历史帧预测轨迹信息对所述至少一个检测目标进行跟踪，获得跟踪结果，包括：创建跟踪器，基于所述跟踪器初始化预测轨迹队列；将所述当前帧预测轨迹信息添加至初始化后的所述预测轨迹队列；判断添加了所述当前帧预测轨迹信息后的所述预测轨迹队列中是否只有所述当前帧预测轨迹信息；若是，给予所述检测目标新的轨迹ID；若否，利用所述跟踪器对所述当前帧检测框与所述预测轨迹队列中的历史帧预测轨迹信息进行匹配，根据匹配结果确定所述检测目标的轨迹ID。In one embodiment, the tracking of the at least one detected target based on the detection frame of the current frame, the predicted trajectory information of the current frame and the predicted trajectory information of the historical frame, and obtaining the tracking result includes: creating a tracker, based on The tracker initializes the predicted trajectory queue; adds the current frame predicted trajectory information to the initialized predicted trajectory queue; judges whether there is only the predicted trajectory queue after adding the current frame predicted trajectory information. Current frame predicted trajectory information; if so, give the detection target a new trajectory ID; if not, use the tracker to match the current frame detection frame with the historical frame predicted trajectory information in the predicted trajectory queue, according to The matching result determines the track ID of the detected target.
在一个实施方式中，所述利用所述跟踪器对所述当前帧检测框与所述预测轨迹队列中的历史帧预测轨迹信息进行匹配，根据匹配结果确定所述检测目标的轨迹ID，包括:将所述当前帧检测框与上一历史帧的第1个轨迹点的预测框进行匹配，其中所述上一历史帧的第1个轨迹点的预测框基于所述上一历史帧预测轨迹信息得到；若匹配成功，则将所述上一历史帧预测轨迹信息对应的轨迹ID作为所述检测目标的轨迹ID；否则，将所述当前帧继续与所述预测轨迹队列中所述上一历史帧之前的其他历史帧逐帧匹配，直到所述当前帧检测框与所述其他历史帧中的一帧历史帧预测轨迹信息匹配成功，则停止匹配，并将匹配成功的所述一帧历史帧预测轨迹信息对应的轨迹ID作为所述检测目标的轨迹ID。In one embodiment, the use of the tracker to match the current frame detection frame with the historical frame predicted trajectory information in the predicted trajectory queue, and determine the trajectory ID of the detection target according to the matching result, including: Match the detection frame of the current frame with the predicted frame of the first trajectory point of the previous historical frame, wherein the predicted frame of the first trajectory point of the previous historical frame is based on the predicted trajectory information of the previous historical frame Obtained; if the matching is successful, then use the track ID corresponding to the predicted track information of the last historical frame as the track ID of the detection target; Other historical frames before the frame are matched frame by frame until the detection frame of the current frame matches the predicted trajectory information of a historical frame in the other historical frames successfully, then the matching is stopped, and the historical frame of the successful matching is The track ID corresponding to the predicted track information is used as the track ID of the detection target.
在一个实施方式中，所述将所述当前帧继续与所述预测轨迹队列中所述上一历史帧之前的其他历史帧逐帧匹配，包括：以当前帧作为第t帧，将所述当前帧检测框与第t-i帧的第i个轨迹点的预测框进行匹配，其中第t-i帧的第i个轨迹点的预测框基于第t-i帧预测轨迹信息得到，其中t和i为正整数，1＜i＜t。In one embodiment, the step of matching the current frame with other historical frames before the last historical frame in the predicted trajectory queue frame by frame includes: taking the current frame as the tth frame, adding the current The frame detection frame is matched with the predicted frame of the i-th track point in the t-i-th frame, where the predicted frame of the i-th track point in the t-i-th frame is obtained based on the predicted track information of the t-i-th frame, where t and i are positive integers, 1 < i < t.
在一个实施方式中，所述利用所述跟踪器对所述当前帧检测框与所述预测轨迹队列中的历史帧预测轨迹信息进行匹配，根据匹配结果确定所述检测目标的轨迹ID，包括：在所述当前帧检测框与所述预测轨迹队列中的所有历史帧预测轨迹信息均匹配失败的情况下，给予所述检测目标新的轨迹ID。In one embodiment, using the tracker to match the detection frame of the current frame with the predicted trajectory information of the historical frame in the predicted trajectory queue, and determining the trajectory ID of the detected target according to the matching result includes: In the case that the detection frame of the current frame fails to match the predicted trajectory information of all historical frames in the predicted trajectory queue, a new trajectory ID is given to the detection target.
在一个实施方式中，所述基于所述跟踪器初始化预测轨迹队列，包括：基于所述跟踪器设置所述预测轨迹队列的队列长度阈值，以及设置预测轨迹队列的时间长度阈值；所述判断添加了所述当前帧预测轨迹信息后的所述预测轨迹队列中是否只有所述当前帧预测轨迹信息之前，所述方法还包括：基于所述时间长度阈值对添加了所述当前帧预测轨迹信息后的所述预测轨迹队列进行有效帧检查，基于检查结果选择性地对所述预测轨迹队列进行更新；和/或基于所述队列长度阈值选择性地对所述预测轨迹队列进行更新。In one embodiment, the initialization of the predicted trajectory queue based on the tracker includes: setting the queue length threshold of the predicted trajectory queue based on the tracker, and setting the time length threshold of the predicted trajectory queue; the judgment adds Whether there is only the predicted trajectory information of the current frame in the predicted trajectory queue after the predicted trajectory information of the current frame is added, the method further includes: after adding the predicted trajectory information of the current frame based on the time length threshold The queue of predicted trajectories is checked for valid frames, and the queue of predicted trajectories is selectively updated based on a check result; and/or the queue of predicted trajectories is selectively updated based on the queue length threshold.
在一个实施方式中，基于所述时间长度阈值对添加了所述当前帧预测轨迹信息后的所述预测轨迹队列进行有效帧检查，基于检查结果选择性地对所述预测轨迹队列进行更新，包括：获取所述预测轨迹队列中的当前帧预测轨迹信息对应的第一时间戳、上一历史帧预测轨迹信息对应的第二时间戳；确定所述第一时间戳与所述第二时间戳的差值；判断所述差值是否小于时间长度阈值；若是，更新所述预测轨迹队列，若否，不更新所述预测轨迹队列。In one embodiment, a valid frame inspection is performed on the predicted trajectory queue after adding the predicted trajectory information of the current frame based on the time length threshold, and selectively updating the predicted trajectory queue based on the inspection result, including : Obtain the first timestamp corresponding to the current frame predicted trajectory information in the predicted trajectory queue, and the second timestamp corresponding to the predicted trajectory information of the last historical frame; determine the relationship between the first timestamp and the second timestamp difference; judging whether the difference is less than a time length threshold; if yes, updating the predicted trajectory queue, if not, not updating the predicted trajectory queue.
在一个实施方式中，所述基于所述队列长度阈值选择性地对所述预测轨迹队列进行更新，包括：在添加了所述当前帧预测轨迹信息后的所述预测轨迹队列的长度超过所述队列长度阈值的情况下，将距离所述当前帧预测轨迹信息最远的历史帧删除。In one embodiment, the selectively updating the predicted trajectory queue based on the queue length threshold includes: the length of the predicted trajectory queue after adding the predicted trajectory information of the current frame exceeds the In the case of the queue length threshold, delete the historical frame farthest from the predicted trajectory information of the current frame.
在一个实施方式中，在利用所述跟踪器对所述当前帧检测框与所述预测轨迹队列中的所述历史帧预测轨迹信息进行匹配之前，所述方法还包括：将所述历史帧预测轨迹信息转换至当前帧对应的车辆坐标系下。In one embodiment, before using the tracker to match the current frame detection frame with the historical frame predicted trajectory information in the predicted trajectory queue, the method further includes: predicting the historical frame The trajectory information is converted to the vehicle coordinate system corresponding to the current frame.
在第二方面，提供一种电子设备，该电子设备包括至少一个处理器和至少一个存储装置，所述存储装置适于存储多条程序代码，所述程序代码适于由所述处理器加载并运行以执行前述任一项所述的目标跟踪方法。In a second aspect, there is provided an electronic device comprising at least one processor and at least one storage device adapted to store pieces of program code adapted to be loaded by the processor and Operate to perform the target tracking method described in any one of the foregoing.
在第三方面，提供一种计算机可读存储介质，该计算机可读存储介质其中存储有多条程序代码，所述程序代码适于由处理器加载并运行以执行前述任一项所述的目标跟踪方法。In a third aspect, there is provided a computer-readable storage medium, wherein the computer-readable storage medium stores a plurality of program codes, and the program codes are adapted to be loaded and run by a processor to perform the target described in any one of the foregoing tracking method.
在第四方面，提供一种车辆，所述车辆包括前述的电子设备。In a fourth aspect, there is provided a vehicle comprising the aforementioned electronic device.
本发明上述一个或多个技术方案，至少具有如下一种或多种有益效果：The above-mentioned one or more technical solutions of the present invention have at least one or more of the following beneficial effects:
本发明中的目标跟踪方法，获取车载传感器采集的传感器数据；将传感器数据输入检测模型，输出至少一个检测目标的当前帧检测框和当前帧预测轨迹信息；基于当前帧检测框、当前帧预测轨迹信息和历史帧预测轨迹信息对至少一个检测目标进行跟踪，获得跟踪结果。如此，利用预测轨迹对目标进行时序跟踪，在保证跟踪速度情况下，提高了目标跟踪的准确性和稳定性。The target tracking method in the present invention obtains the sensor data collected by the vehicle-mounted sensor; inputs the sensor data into the detection model, and outputs at least one current frame detection frame and current frame prediction track information of the detection target; based on the current frame detection frame and the current frame prediction track information Information and historical frame prediction trajectory information are used to track at least one detection target to obtain a tracking result. In this way, using the predicted trajectory to track the target in time series improves the accuracy and stability of the target tracking under the condition of ensuring the tracking speed.
附图说明Description of drawings
参照附图，本发明的公开内容将变得更易理解。本领域技术人员容易理解的是：这些附图仅仅用于说明的目的，而并非意在对本发明的保护范围组成限制。此外，图中类似的数字用以表示类似的部件，其中：The disclosure of the present invention will become more comprehensible with reference to the accompanying drawings. Those skilled in the art can easily understand that: these drawings are only for the purpose of illustration, and are not intended to limit the protection scope of the present invention. In addition, like numerals are used to designate like parts in the drawings, wherein:
图1是根据本发明的一个实施例的目标跟踪方法的主要步骤流程示意图；Fig. 1 is a schematic flow chart of the main steps of a target tracking method according to an embodiment of the present invention;
图2是一个实施例中对至少一个检测目标进行跟踪的流程示意图；Fig. 2 is a schematic flow chart of tracking at least one detection target in an embodiment;
图3是一个实施例中目标跟踪方法的完整流程示意图；Fig. 3 is a complete schematic flow chart of a target tracking method in an embodiment;
图4是一个实施例中电子设备的结构示意图。Fig. 4 is a schematic structural diagram of an electronic device in an embodiment.
具体实施方式Detailed ways
下面参照附图来描述本发明的一些实施方式。本领域技术人员应当理解的是，这些实施方式仅仅用于解释本发明的技术原理，并非旨在限制本发明的保护范围。Some embodiments of the present invention are described below with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are only used to explain the technical principles of the present invention, and are not intended to limit the protection scope of the present invention.
在本发明的描述中，“模块”、“处理器”可以包括硬件、软件或者两者的组合。一个模块可以包括硬件电路，各种合适的感应器，通信端口，存储器，也可以包括软件部分，比如程序代码，也可以是软件和硬件的组合。处理器可以是中央处理器、微处理器、图像处理器、数字信号处理器或者其他任何合适的处理器。处理器具有数据和/或信号处理功能。处理器可以以软件方式实现、硬件方式实现或者二者结合方式实现。非暂时性的计算机可读存储介质包括任何合适的可存储程序代码的介质，比如磁碟、硬盘、光碟、闪存、只读存储器、随机存取存储器等等。术语“A和/或B”表示所有可能的A与B的组合，比如只是A、只是B或者A和B。术语“至少一个A或B”或者“A和B中的至少一个”含义与“A和/或B”类似，可以包括只是A、只是B或者A和B。单数形式的术语“一个”、“这个”也可以包含复数形式。In the description of the present invention, "module" and "processor" may include hardware, software or a combination of both. A module may include hardware circuits, various suitable sensors, communication ports, memory, and may also include software parts, such as program codes, or a combination of software and hardware. The processor may be a central processing unit, a microprocessor, an image processor, a digital signal processor or any other suitable processor. The processor has data and/or signal processing functions. The processor can be implemented in software, hardware or a combination of both. The non-transitory computer readable storage medium includes any suitable medium that can store program code, such as magnetic disks, hard disks, optical disks, flash memory, read only memory, random access memory, and the like. The term "A and/or B" means all possible combinations of A and B, such as only A, only B or A and B. The term "at least one of A or B" or "at least one of A and B" has a similar meaning to "A and/or B" and may include only A, only B or both A and B. The terms "a" and "the" in the singular may also include plural forms.
目前传统的目标跟踪方法大多为光流法、滤波类算法，其中光流法受环境影响较大，在复杂环境下跟踪的准确度较差，滤波类算法对于目标形态变化、遮挡或目标消失等情况下跟踪的准确度较差。At present, most of the traditional target tracking methods are optical flow method and filtering algorithm. Among them, the optical flow method is greatly affected by the environment, and the tracking accuracy in complex environments is poor. The tracking accuracy is poor.
为此，本申请提出了一种目标跟踪方法、电子设备、存储介质及车辆，获取车载传感器采集的传感器数据；将传感器数据输入检测模型，输出至少一个检测目标的当前帧检测框和当前帧预测轨迹信息；基于当前帧检测框、当前帧预测轨迹信息和历史帧预测轨迹信息对至少一个检测目标进行跟踪，获得跟踪结果。如此，利用预测轨迹对目标进行时序跟踪，在保证跟踪速度情况下，提高了目标跟踪的准确性和稳定性。For this reason, the present application proposes a target tracking method, electronic equipment, storage media and vehicles to obtain sensor data collected by on-board sensors; input the sensor data into the detection model, and output the current frame detection frame and current frame prediction of at least one detection target Trajectory information: based on the current frame detection frame, the current frame predicted trajectory information and the historical frame predicted trajectory information, at least one detection target is tracked to obtain a tracking result. In this way, using the predicted trajectory to track the target in time series improves the accuracy and stability of the target tracking under the condition of ensuring the tracking speed.
参阅附图1，图1是根据本发明的一个实施例的目标跟踪方法的主要步骤流程示意图。Referring to accompanying
如图1所示，本发明实施例中的目标跟踪方法主要包括下列步骤S101-步骤S103。As shown in FIG. 1 , the target tracking method in the embodiment of the present invention mainly includes the following steps S101 - S103.
步骤S101：获取车载传感器采集的传感器数据。Step S101: Obtain sensor data collected by vehicle sensors.
步骤S102：将传感器数据输入网络模型，输出至少一个检测目标的当前帧检测框和当前帧预测轨迹信息。Step S102: Input the sensor data into the network model, and output the current frame detection frame and current frame predicted trajectory information of at least one detected target.
步骤S103：基于当前帧检测框、当前帧预测轨迹信息和历史帧预测轨迹信息对至少一个检测目标进行跟踪，获得跟踪结果。Step S103: Track at least one detection target based on the detection frame of the current frame, the predicted trajectory information of the current frame and the predicted trajectory information of the historical frame, and obtain a tracking result.
基于上述步骤S101-步骤S103，获取车载传感器采集的传感器数据；将传感器数据输入检测模型，输出至少一个检测目标的当前帧检测框和当前帧预测轨迹信息；基于当前帧检测框、当前帧预测轨迹信息和历史帧预测轨迹信息对至少一个检测目标进行跟踪，获得跟踪结果。如此，利用预测轨迹对目标进行时序跟踪，在保证跟踪速度情况下，提高了目标跟踪的准确性和稳定性。Based on the above step S101-step S103, obtain the sensor data collected by the vehicle sensor; input the sensor data into the detection model, and output the current frame detection frame and current frame prediction trajectory information of at least one detection target; based on the current frame detection frame and current frame prediction trajectory information Information and historical frame prediction trajectory information are used to track at least one detection target to obtain a tracking result. In this way, using the predicted trajectory to track the target in time series improves the accuracy and stability of the target tracking under the condition of ensuring the tracking speed.
下面分别对上述步骤S101至步骤S103作进一步说明。The above step S101 to step S103 will be further described below respectively.
在步骤S101中，车载传感器可以是摄像头和激光雷达中的任意一种。车载传感器采集的传感器数据可以是视频帧图像，也可以是激光雷达采集的点云图像。In step S101, the vehicle sensor may be any one of a camera and a laser radar. The sensor data collected by the vehicle sensor can be a video frame image or a point cloud image collected by the lidar.
另外，获取的车载传感器采集的传感器图像可以是一帧图像，也可以是多帧图像，例如不同视角传感器采集的传感器数据。输入的信息包含多帧数据，能够提升检测效果的稳定性和预测轨迹的准确性，但是多帧数据的跟踪对于时间成本和算力的要求较高。在本实施例中，优选获取一帧传感器采集数据。In addition, the acquired sensor image collected by the vehicle-mounted sensor may be one frame of image or multiple frames of images, for example, sensor data collected by sensors with different viewing angles. The input information contains multi-frame data, which can improve the stability of the detection effect and the accuracy of the predicted trajectory, but the tracking of multi-frame data requires high time cost and computing power. In this embodiment, preferably, one frame of sensor acquisition data is acquired.
以上是对步骤S101的进一步说明，下面继续对步骤S102作进一步说明。The above is a further description of step S101, and further description of step S102 will be continued below.
网络模型是至少包含了检测和预测两种任务的端到端的网络模型。将传感器数据输入该网络模型，经过主干网络特征提取、多任务的检测头和轨迹预测头，输出所述传感器数据中对应的至少一个检测目标的当前帧检测框和当前帧预测轨迹信息。其中预测轨迹信息包括未来时刻的轨迹点坐标和朝向角等。The network model is an end-to-end network model that includes at least two tasks of detection and prediction. Input the sensor data into the network model, and output the current frame detection frame and the current frame prediction trajectory information of at least one detection target corresponding to the sensor data through backbone network feature extraction, multi-task detection head and trajectory prediction head. The predicted trajectory information includes trajectory point coordinates and orientation angles at a future moment.
IntentNet可以作为所述网络模型的一个示例，但不限于此。IntentNet can be used as an example of the network model, but is not limited thereto.
以上是对步骤S102的进一步说明，下面继续对步骤S103作进一步说明。The above is a further description of step S102, and further description of step S103 will be continued below.
具体地，上述步骤S103可通过下述步骤S1031至步骤S1035实现。Specifically, the above step S103 may be implemented through the following steps S1031 to S1035.
步骤S1031：创建跟踪器，基于所述跟踪器初始化预测轨迹队列。Step S1031: Create a tracker, and initialize a predicted trajectory queue based on the tracker.
跟踪器是一个对象，创建跟踪器的过程就是新建一个跟踪器对象，同时初始化的过程。A tracker is an object, and the process of creating a tracker is to create a new tracker object and initialize it at the same time.
预测轨迹队列用于存储当前帧信息和历史帧信息，例如当前帧预测轨迹信息、历史帧预测轨迹信息和历史帧检测框等。The predicted trajectory queue is used to store current frame information and historical frame information, such as current frame predicted trajectory information, historical frame predicted trajectory information, and historical frame detection boxes.
在一个具体实施例中，所述基于所述跟踪器初始化预测轨迹队列，包括：基于所述跟踪器设置所述预测轨迹队列的队列长度阈值，以及设置预测轨迹队列的时间长度阈值；所述判断添加了所述当前帧预测轨迹信息后的所述预测轨迹队列中是否只有所述当前帧预测轨迹信息之前，所述方法还包括：基于所述时间长度阈值对添加了所述当前帧预测轨迹信息后的所述预测轨迹队列进行有效帧检查，基于检查结果选择性地对所述预测轨迹队列进行更新；和/或基于所述队列长度阈值选择性地对所述预测轨迹队列进行更新。In a specific embodiment, the initialization of the predicted trajectory queue based on the tracker includes: setting the queue length threshold of the predicted trajectory queue based on the tracker, and setting the time length threshold of the predicted trajectory queue; the judgment Whether there is only the predicted trajectory information of the current frame in the predicted trajectory queue after adding the predicted trajectory information of the current frame, the method also includes: adding the predicted trajectory information of the current frame based on the time length threshold The subsequent queue of predicted trajectories is checked for valid frames, and the queue of predicted trajectories is selectively updated based on the check result; and/or the queue of predicted trajectories is selectively updated based on the queue length threshold.
更新预测轨迹队列可以是对不满足有效帧检查或者不满足队列长度阈值的历史帧信息删除，或者从历史帧信息集合中获取新的历史帧信息存储至所述预测轨迹队列。Updating the predicted trajectory queue may be to delete the historical frame information that does not meet the valid frame check or the queue length threshold, or obtain new historical frame information from the historical frame information set and store it in the predicted trajectory queue.
在一个具体实施例中，基于所述时间长度阈值对添加了所述当前帧预测轨迹信息后的所述预测轨迹队列进行有效帧检查，基于检查结果选择性地对所述预测轨迹队列进行更新，包括：获取所述预测轨迹队列中的当前帧预测轨迹信息对应的第一时间戳、上一历史帧预测轨迹信息对应的第二时间戳；确定所述第一时间戳与所述第二时间戳的差值；判断所述差值是否小于时间长度阈值；若是，更新所述预测轨迹队列，若否，不更新所述预测轨迹队列。In a specific embodiment, a valid frame inspection is performed on the predicted trajectory queue after adding the predicted trajectory information of the current frame based on the time length threshold, and the predicted trajectory queue is selectively updated based on the inspection result, Including: obtaining the first timestamp corresponding to the current frame predicted trajectory information in the predicted trajectory queue, and the second timestamp corresponding to the predicted trajectory information of the previous historical frame; determining the first timestamp and the second timestamp Determine whether the difference is less than the time length threshold; if yes, update the predicted trajectory queue, if not, do not update the predicted trajectory queue.
具体来说，每一帧信息都有对应的时间戳，对预测轨迹队列进行有效帧检查指的是检查第t帧（当前帧）时间戳与第t-1帧的时间戳的差值是否位于合理区间内。具体地，当前帧预测轨迹信息对应第一时间戳，上一历史帧预测轨迹信息对应第二时间戳，根据第一时间戳和第二时间戳计算两者的差值，判断差值是否位于时间长度阈值范围内；若是，更新预测轨迹队列，若否，不更新预测轨迹队列。Specifically, each frame of information has a corresponding timestamp, and checking the valid frame of the predicted trajectory queue refers to checking whether the difference between the timestamp of the tth frame (current frame) and the timestamp of the t-1th frame is in within a reasonable range. Specifically, the predicted trajectory information of the current frame corresponds to the first timestamp, and the predicted trajectory information of the last historical frame corresponds to the second timestamp. The difference between the two is calculated according to the first timestamp and the second timestamp, and it is judged whether the difference is at the time The length is within the threshold range; if yes, update the predicted trajectory queue, if not, do not update the predicted trajectory queue.
通过对预测轨迹队列进行有效帧检查，能够保证上一历史帧时间戳与当前帧时间戳的差值始终位于合理的时间范围，从而提高对目标跟踪的准确性。By checking the effective frame of the predicted trajectory queue, it can ensure that the difference between the timestamp of the previous historical frame and the timestamp of the current frame is always within a reasonable time range, thereby improving the accuracy of target tracking.
在一个具体实施例中，所述基于所述队列长度阈值选择性地对所述预测轨迹队列进行更新，包括：在添加了所述当前帧预测轨迹信息后的所述预测轨迹队列的长度超过所述队列长度阈值的情况下，将距离所述当前帧预测轨迹信息最远的历史帧删除。In a specific embodiment, the selectively updating the predicted trajectory queue based on the queue length threshold includes: the length of the predicted trajectory queue after adding the predicted trajectory information of the current frame exceeds the specified In the case of the above-mentioned queue length threshold, the historical frame farthest from the predicted trajectory information of the current frame is deleted.
在维护的预测轨迹队列的长度超过队列长度阈值的情况下，将距离所述当前帧信息最远的历史帧删除。If the length of the maintained predicted trajectory queue exceeds the queue length threshold, the historical frame farthest from the current frame information is deleted.
队列长度阈值用于维护预测轨迹队列，维护一定长度的预测轨迹队列，可以提升遮挡或者物体消失状况下的跟踪效果。The queue length threshold is used to maintain the predicted trajectory queue. Maintaining a certain length of predicted trajectory queue can improve the tracking effect under the condition of occlusion or object disappearance.
步骤S1032：将所述当前帧预测轨迹信息添加至初始化后的所述预测轨迹队列。Step S1032: Add the predicted trajectory information of the current frame to the initialized predicted trajectory queue.
在一个实施例中，当前帧的检测框信息也可以添加至初始化后的预测轨迹队列。也可以重新构建一个存储检测框的检测框队列，以将所述当前帧检测框存储至所述检测框队列中。In one embodiment, the detection frame information of the current frame may also be added to the initialized predicted trajectory queue. A detection frame queue for storing detection frames may also be reconstructed, so as to store the detection frame of the current frame into the detection frame queue.
步骤S1033：判断添加了所述当前帧预测轨迹信息后的所述预测轨迹队列中是否只有所述当前帧预测轨迹信息。Step S1033: Determine whether there is only the predicted trajectory information of the current frame in the predicted trajectory queue after adding the predicted trajectory information of the current frame.
通常来说，进行初始跟踪时，预测轨迹队列中只有当前帧预测轨迹信息。后续过程中对目标对象的跟踪，预测轨迹队列中同时存在当前帧预测轨迹信息和历史帧预测轨迹信息。Generally speaking, when initial tracking is performed, only the predicted trajectory information of the current frame is in the predicted trajectory queue. For the tracking of the target object in the subsequent process, the predicted trajectory information of the current frame and the predicted trajectory information of the historical frame exist in the predicted trajectory queue at the same time.
步骤S1034：若是，给予所述检测目标新的轨迹ID。Step S1034: If yes, give the detected object a new track ID.
步骤S1035：若否，利用跟踪器对当前帧检测框与预测轨迹队列中的历史帧预测轨迹信息进行匹配，根据匹配结果确定检测目标的轨迹ID。Step S1035: If not, use the tracker to match the detection frame of the current frame with the predicted trajectory information of the historical frame in the predicted trajectory queue, and determine the trajectory ID of the detection target according to the matching result.
轨迹ID（track ID），具体指目标跟踪ID。Track ID (track ID), specifically refers to the target tracking ID.
具体来说，当前帧检测框与预测轨迹队列中的历史帧预测轨迹信息的匹配，主要是将当前帧检测框与每一帧历史帧预测轨迹信息对应的预测框进行逐帧匹配，从而确定预测轨迹队列中是否存在与当前帧检测框能够匹配成功的历史帧。Specifically, the matching of the current frame detection frame and the historical frame prediction trajectory information in the prediction trajectory queue is mainly to match the current frame detection frame with the prediction frame corresponding to each frame of historical frame prediction trajectory information frame by frame, so as to determine the prediction Whether there is a historical frame that can successfully match the current frame detection box in the track queue.
在一个具体实施方式中，所述利用所述跟踪器对所述当前帧检测框与所述预测轨迹队列中的历史帧预测轨迹信息进行匹配，根据匹配结果确定所述检测目标的轨迹ID，包括:将所述当前帧检测框与上一历史帧的第1个轨迹点的预测框进行匹配，其中所述上一历史帧的第1个轨迹点的预测框基于所述上一历史帧预测轨迹信息得到；若匹配成功，则将所述上一历史帧预测轨迹信息对应的轨迹ID作为所述检测目标的轨迹ID；否则，将所述当前帧继续与所述预测轨迹队列中所述上一历史帧之前的其他历史帧逐帧匹配，直到所述当前帧检测框与所述其他历史帧中的一帧历史帧预测轨迹信息匹配成功，则停止匹配，并将匹配成功的所述一帧历史帧预测轨迹信息对应的轨迹ID作为所述检测目标的轨迹ID。In a specific embodiment, the use of the tracker to match the detection frame of the current frame with the predicted trajectory information of the historical frame in the predicted trajectory queue, and determine the trajectory ID of the detected target according to the matching result, including : matching the current frame detection frame with the prediction frame of the first trajectory point of the last historical frame, wherein the prediction frame of the first trajectory point of the last historical frame is based on the predicted trajectory of the last historical frame Information is obtained; if the matching is successful, the track ID corresponding to the predicted track information of the last historical frame is used as the track ID of the detection target; Other historical frames before the historical frame are matched frame by frame until the detection frame of the current frame matches the predicted track information of a historical frame in the other historical frames successfully, then the matching is stopped, and the historical frame of the successfully matched frame is The track ID corresponding to the frame prediction track information is used as the track ID of the detection target.
每帧历史帧预测轨迹信息包括多个轨迹点以及每个轨迹点的朝向。每个轨迹点、轨迹点的朝向结合该帧的检测框即可得到每帧每个轨迹点对应的预测框。The predicted trajectory information of each frame history frame includes multiple trajectory points and the orientation of each trajectory point. Each track point and the orientation of the track point are combined with the detection frame of the frame to obtain the prediction frame corresponding to each track point in each frame.
具体来说，首先将当前帧的检测框与上一历史帧的第1个轨迹点的预测框进行匹配，若匹配成功，则将上一历史帧预测轨迹信息对应的轨迹ID作为检测目标的轨迹ID；否则，将当前帧继续与预测轨迹队列中上一历史帧之前的其他历史帧逐帧匹配，直到当前帧检测框与其他历史帧中的一帧历史帧预测轨迹信息匹配成功，则停止匹配，并将匹配成功的一帧历史帧预测轨迹信息对应的轨迹ID作为检测目标的轨迹ID。Specifically, firstly, the detection frame of the current frame is matched with the prediction frame of the first trajectory point of the previous historical frame, and if the match is successful, the trajectory ID corresponding to the predicted trajectory information of the previous historical frame is used as the trajectory of the detection target ID; otherwise, continue to match the current frame with other historical frames before the previous historical frame in the predicted trajectory queue frame by frame, until the current frame detection frame successfully matches the predicted trajectory information of a historical frame in other historical frames, then stop matching , and use the track ID corresponding to the predicted track information of one frame of historical frame that matches successfully as the track ID of the detection target.
当上一历史帧预测轨迹信息有多个时，此时有多个预测框，需要将检测框与多个预测框分别进行匹配，具体计算检测框与每个预测框之间的交并比Iou，从多个交并比中选出符合条件的至少一个交并比，然后使用匈牙利匹配算法确定与所述检测框能够匹配成功的预测框。When there are multiple predicted trajectory information in the previous historical frame, there are multiple prediction frames at this time, and the detection frame needs to be matched with multiple prediction frames, and the intersection ratio Iou between the detection frame and each prediction frame is specifically calculated , select at least one intersection and union ratio that meets the conditions from multiple intersection and union ratios, and then use the Hungarian matching algorithm to determine the prediction frame that can be successfully matched with the detection frame.
在一个具体实施方式中，所述将所述当前帧继续与所述预测轨迹队列中所述上一历史帧之前的其他历史帧逐帧匹配，包括：以当前帧作为第t帧，将所述当前帧检测框与第t-i帧的第i个轨迹点的预测框进行匹配，其中第t-i帧的第i个轨迹点的预测框基于第t-i帧预测轨迹信息得到，其中t和i为正整数，1＜i＜t。In a specific implementation manner, the matching of the current frame with other historical frames before the last historical frame in the predicted trajectory queue frame by frame includes: taking the current frame as the tth frame, adding the The detection frame of the current frame is matched with the predicted frame of the i-th track point of the t-i-th frame, wherein the predicted frame of the i-th track point of the t-i-th frame is obtained based on the predicted track information of the t-i-th frame, where t and i are positive integers, 1<i<t.
具体来说，如图2所示，在当前帧检测框（第t帧）与上一历史帧（第t-1帧）匹配失败的情况下，继续将当前帧检测框与第t-2帧的第二个轨迹点的预测框进行匹配，若匹配成功，则将第t-2帧预测轨迹信息对应的轨迹ID作为检测目标的轨迹ID。若匹配失败，继续将当前帧检测框与第t-3帧进行匹配，重复执行该步骤，直至当前帧检测框与预测轨迹队列中的历史帧匹配成功，输出检测目标的轨迹ID。Specifically, as shown in Figure 2, when the current frame detection frame (t-th frame) fails to match the previous historical frame (t-1th frame), continue to match the current frame detection frame with the t-2th frame Match the prediction frame of the second trajectory point in the frame. If the match is successful, the trajectory ID corresponding to the predicted trajectory information of the t-2th frame is used as the trajectory ID of the detection target. If the matching fails, continue to match the detection frame of the current frame with the t-3th frame, and repeat this step until the detection frame of the current frame is successfully matched with the historical frames in the predicted trajectory queue, and output the trajectory ID of the detection target.
在一个具体实施方式中，所述利用所述跟踪器对所述当前帧检测框与所述预测轨迹队列中的历史帧预测轨迹信息进行匹配，根据匹配结果确定所述检测目标的轨迹ID，包括：在所述当前帧检测框与所述预测轨迹队列中的所有历史帧预测轨迹信息均匹配失败的情况下，给予所述检测目标新的轨迹ID。In a specific embodiment, the use of the tracker to match the detection frame of the current frame with the predicted trajectory information of the historical frame in the predicted trajectory queue, and determine the trajectory ID of the detected target according to the matching result, including : In the case that the detection frame of the current frame fails to match all the predicted trajectory information of the historical frames in the predicted trajectory queue, give the detection target a new trajectory ID.
具体来说，在当前帧检测框与预测轨迹队列中的所有历史帧均匹配失败的情况下，给予检测目标新的轨迹ID，说明该检测目标是新出现的目标。Specifically, in the case that the detection frame of the current frame fails to match all the historical frames in the predicted trajectory queue, a new trajectory ID is given to the detected object, indicating that the detected object is a new object.
通过维护预测轨迹队列，并将当前帧信息与历史预测信息进行匹配，获得跟踪目标，一定程度上解决了跟踪过程中检测目标的形态变化、遮挡、消失等问题，从而输出稳定的跟踪目标。By maintaining the predicted trajectory queue and matching the current frame information with the historical predicted information to obtain the tracking target, it solves the problems of shape change, occlusion, and disappearance of the detection target during the tracking process to a certain extent, thereby outputting a stable tracking target.
在一个实施例中，具体如图3所示，在构建跟踪器之后、以及目标跟踪之前，还可以对步骤S102输出至少一个检测目标的当前帧检测框和当前帧预测轨迹信息进行数据预处理。示例性地，检测框信息和预测轨迹信息中包含了框的类别和置信度，设置不同类别的置信度阈值，过滤掉置信度较小的检测框和预测轨迹。另外，轨迹的多模态预测，即一个框预测多条轨迹的情况，需要筛选出置信度最大的预测轨迹。如此，减小了目标跟踪的复杂度，提高了目标跟踪的效率和准确度。In one embodiment, as specifically shown in FIG. 3 , after building the tracker and before tracking the target, data preprocessing may be performed on the current frame detection frame and current frame predicted trajectory information of at least one detected target output in step S102 . Exemplarily, the detection frame information and the predicted track information include the category and confidence of the frame, and the confidence thresholds of different categories are set to filter out the detection frames and predicted tracks with lower confidence. In addition, the multimodal prediction of trajectories, that is, the case where one box predicts multiple trajectories, needs to filter out the predicted trajectory with the highest confidence. In this way, the complexity of target tracking is reduced, and the efficiency and accuracy of target tracking are improved.
在一个具体实施方式中，在利用所述跟踪器对所述当前帧检测框与所述预测轨迹队列中的所述历史帧预测轨迹信息进行匹配之前，所述方法还包括：将所述历史帧预测轨迹信息转换至当前帧对应的车辆坐标系下。In a specific embodiment, before using the tracker to match the detection frame of the current frame with the predicted trajectory information of the historical frame in the predicted trajectory queue, the method further includes: The predicted trajectory information is converted to the vehicle coordinate system corresponding to the current frame.
具体的，在匹配之前，还可以根据自车定位信息确定历史帧信息转换至当前帧的坐标变换矩阵，进而根据坐标变换矩阵将历史帧预测轨迹信息转换至当前帧对应的车辆坐标系下。如此，方便后续对当前帧检测框与历史帧信息的匹配。Specifically, before matching, it is also possible to determine the coordinate transformation matrix for converting the historical frame information to the current frame according to the vehicle positioning information, and then convert the predicted trajectory information of the historical frame to the vehicle coordinate system corresponding to the current frame according to the coordinate transformation matrix. In this way, it is convenient to subsequently match the current frame detection frame with historical frame information.
需要指出的是，尽管上述实施例中将各个步骤按照特定的先后顺序进行了描述，但是本领域技术人员可以理解，为了实现本发明的效果，不同的步骤之间并非必须按照这样的顺序执行，其可以同时（并行）执行或以其他顺序执行，这些变化都在本发明的保护范围之内。It should be pointed out that, although the steps are described in a specific order in the above embodiments, those skilled in the art can understand that in order to achieve the effect of the present invention, different steps do not have to be executed in this order. They can be executed simultaneously (in parallel) or in other sequences, and these variations are within the protection scope of the present invention.
本领域技术人员能够理解的是，本发明实现上述一实施例的方法中的全部或部分流程，也可以通过计算机程序来指令相关的硬件来完成，所述的计算机程序可存储于一计算机可读存储介质中，该计算机程序在被处理器执行时，可实现上述各个方法实施例的步骤。其中，所述计算机程序包括计算机程序代码，所述计算机程序代码可以为源代码形式、对象代码形式、可执行文件或某些中间形式等。所述计算机可读存储介质可以包括：能够携带所述计算机程序代码的任何实体或装置、介质、U盘、移动硬盘、磁碟、光盘、计算机存储器、只读存储器、随机存取存储器、电载波信号、电信信号以及软件分发介质等。需要说明的是，所述计算机可读存储介质包含的内容可以根据司法管辖区内立法和专利实践的要求进行适当的增减，例如在某些司法管辖区，根据立法和专利实践，计算机可读存储介质不包括电载波信号和电信信号。Those skilled in the art can understand that all or part of the process in the method of the above-mentioned embodiment of the present invention can also be completed by instructing related hardware through a computer program, and the computer program can be stored in a computer-readable In the storage medium, when the computer program is executed by the processor, the steps of the above-mentioned various method embodiments can be realized. Wherein, the computer program includes computer program code, and the computer program code may be in the form of source code, object code, executable file or some intermediate form. The computer-readable storage medium may include: any entity or device capable of carrying the computer program code, medium, U disk, removable hard disk, magnetic disk, optical disk, computer memory, read-only memory, random access memory, electric carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content contained in the computer-readable storage medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable Storage media excludes electrical carrier signals and telecommunication signals.
进一步，本发明还提供了一种电子设备。在根据本发明的一个电子设备实施例中，具体如图4所示，电子设备包括至少一个处理器41和至少一个存储装置42，存储装置可以被配置成存储执行上述方法实施例的目标跟踪方法的程序，处理器可以被配置成用于执行存储装置中的程序，该程序包括但不限于执行上述方法实施例的目标跟踪方法的程序。为了便于说明，仅示出了与本发明实施例相关的部分，具体技术细节未揭示的，请参照本发明实施例方法部分。Further, the present invention also provides an electronic device. In an embodiment of an electronic device according to the present invention, as specifically shown in FIG. 4 , the electronic device includes at least one
在本发明实施例中电子设备可以是包括各种设备形成的控制装置设备。在一些可能的实施方式中，电子设备可以包括多个存储装置和多个处理器。而执行上述方法实施例的目标跟踪方法的程序可以被分割成多段子程序，每段子程序分别可以由处理器加载并运行以执行上述方法实施例的目标跟踪方法的不同步骤。具体地，每段子程序可以分别存储在不同的存储装置中，每个处理器可以被配置成用于执行一个或多个存储装置中的程序，以共同实现上述方法实施例的目标跟踪方法，即每个处理器分别执行上述方法实施例的目标跟踪方法的不同步骤，来共同实现上述方法实施例的目标跟踪方法。In the embodiment of the present invention, the electronic device may be a control device device formed of various devices. In some possible implementation manners, an electronic device may include multiple storage devices and multiple processors. The program for executing the object tracking method of the above method embodiment can be divided into multiple subroutines, and each subroutine can be loaded and run by a processor to execute different steps of the object tracking method of the above method embodiment. Specifically, each subroutine can be stored in a different storage device, and each processor can be configured to execute the programs in one or more storage devices, so as to jointly implement the target tracking method of the above method embodiment, namely Each processor respectively executes different steps of the object tracking method of the above method embodiment to jointly implement the object tracking method of the above method embodiment.
上述多个处理器可以是部署于同一个设备上的处理器，例如上述电子设备可以是由多个处理器组成的高性能设备，上述多个处理器可以是该高性能设备上配置的处理器。此外，上述多个处理器也可以是部署于不同设备上的处理器，例如上述电子设备可以是服务器集群，上述多个处理器可以是服务器集群中不同服务器上的处理器。The above-mentioned multiple processors may be processors deployed on the same device, for example, the above-mentioned electronic device may be a high-performance device composed of multiple processors, and the above-mentioned multiple processors may be processors configured on the high-performance device . In addition, the above multiple processors may also be processors deployed on different devices, for example, the above electronic device may be a server cluster, and the above multiple processors may be processors on different servers in the server cluster.
进一步，本发明还提供了一种计算机可读存储介质。在根据本发明的一个计算机可读存储介质实施例中，计算机可读存储介质可以被配置成存储执行上述方法实施例的目标跟踪方法的程序，该程序可以由处理器加载并运行以实现上述目标跟踪方法。为了便于说明，仅示出了与本发明实施例相关的部分，具体技术细节未揭示的，请参照本发明实施例方法部分。该计算机可读存储介质可以是包括各种电子设备形成的存储装置设备，可选的，本发明实施例中计算机可读存储介质是非暂时性的计算机可读存储介质。Further, the present invention also provides a computer-readable storage medium. In an embodiment of a computer-readable storage medium according to the present invention, the computer-readable storage medium may be configured to store a program for executing the target tracking method of the above-mentioned method embodiment, and the program may be loaded and run by a processor to achieve the above-mentioned target tracking method. For ease of description, only the parts related to the embodiments of the present invention are shown, and for specific technical details not disclosed, please refer to the method part of the embodiments of the present invention. The computer-readable storage medium may be a storage device formed by various electronic devices. Optionally, the computer-readable storage medium in this embodiment of the present invention is a non-transitory computer-readable storage medium.
进一步，本发明还提供了一种车辆，所述车辆包括前述的电子设备。Further, the present invention also provides a vehicle, the vehicle includes the aforementioned electronic equipment.
至此，已经结合附图所示的优选实施方式描述了本发明的技术方案，但是，本领域技术人员容易理解的是，本发明的保护范围显然不局限于这些具体实施方式。在不偏离本发明的原理的前提下，本领域技术人员可以对相关技术特征作出等同的更改或替换，这些更改或替换之后的技术方案都将落入本发明的保护范围之内。So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings, but those skilled in the art will easily understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to relevant technical features, and the technical solutions after these changes or substitutions will all fall within the protection scope of the present invention.
Claims (12)
Translated fromChinesePriority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310171954.8ACN115965657B (en) | 2023-02-28 | 2023-02-28 | Target tracking method, electronic device, storage medium and vehicle |
| PCT/CN2023/140010WO2024179141A1 (en) | 2023-02-28 | 2023-12-19 | Object tracking method, electronic device, storage medium and vehicle |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310171954.8ACN115965657B (en) | 2023-02-28 | 2023-02-28 | Target tracking method, electronic device, storage medium and vehicle |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115965657A CN115965657A (en) | 2023-04-14 |
| CN115965657Btrue CN115965657B (en) | 2023-06-02 |
Family
ID=85894647
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310171954.8AActiveCN115965657B (en) | 2023-02-28 | 2023-02-28 | Target tracking method, electronic device, storage medium and vehicle |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN115965657B (en) |
| WO (1) | WO2024179141A1 (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115965657B (en)* | 2023-02-28 | 2023-06-02 | 安徽蔚来智驾科技有限公司 | Target tracking method, electronic device, storage medium and vehicle |
| CN117292306A (en)* | 2023-11-27 | 2023-12-26 | 四川迪晟新达类脑智能技术有限公司 | Edge equipment-oriented vehicle target detection optimization method and device |
| CN117351039B (en)* | 2023-12-06 | 2024-02-02 | 广州紫为云科技有限公司 | Nonlinear multi-target tracking method based on feature query |
| CN117830356B (en)* | 2023-12-29 | 2025-08-26 | 苏州万店掌网络科技有限公司 | Target tracking method, device, equipment and medium |
| CN118279350B (en)* | 2024-06-03 | 2024-08-09 | 杭州登虹科技有限公司 | Target tracking method, device, electronic equipment and computer readable storage medium based on end-to-end model |
| CN119810505B (en)* | 2024-12-09 | 2025-09-23 | 东南大学 | Method for smoothing target detection label based on historical information |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102065275A (en)* | 2009-11-17 | 2011-05-18 | 中国科学院电子学研究所 | Multi-target tracking method in intelligent video monitoring system |
| WO2021134172A1 (en)* | 2019-12-30 | 2021-07-08 | 华为技术有限公司 | Trajectory prediction method and related device |
| CN114169241A (en)* | 2021-12-09 | 2022-03-11 | 北京邮电大学 | An end-to-end multi-target recognition, tracking and prediction method |
| WO2022127180A1 (en)* | 2020-12-17 | 2022-06-23 | 深圳云天励飞技术股份有限公司 | Target tracking method and apparatus, and electronic device and storage medium |
| WO2022142417A1 (en)* | 2020-12-31 | 2022-07-07 | 深圳云天励飞技术股份有限公司 | Target tracking method and apparatus, electronic device, and storage medium |
| WO2022142918A1 (en)* | 2020-12-31 | 2022-07-07 | 深圳云天励飞技术股份有限公司 | Multi-target tracking system and method |
| CN115311330A (en)* | 2019-10-11 | 2022-11-08 | 杭州云栖智慧视通科技有限公司 | Video multi-target tracking method based on position prediction |
Family Cites Families (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7298867B2 (en)* | 2004-02-20 | 2007-11-20 | Lockheed Martin Corporation | Component association tracker system and method |
| US9255989B2 (en)* | 2012-07-24 | 2016-02-09 | Toyota Motor Engineering & Manufacturing North America, Inc. | Tracking on-road vehicles with sensors of different modalities |
| US10482572B2 (en)* | 2017-10-06 | 2019-11-19 | Ford Global Technologies, Llc | Fusion of motion and appearance features for object detection and trajectory prediction |
| CN110751674A (en)* | 2018-07-24 | 2020-02-04 | 北京深鉴智能科技有限公司 | Multi-target tracking method and corresponding video analysis system |
| CN111428943B (en)* | 2020-04-23 | 2021-08-03 | 福瑞泰克智能系统有限公司 | Method, device and computer device for predicting obstacle vehicle track |
| CN111986228B (en)* | 2020-09-02 | 2023-06-02 | 华侨大学 | A pedestrian tracking method, device and medium based on LSTM model in escalator scene |
| CN112200830A (en)* | 2020-09-11 | 2021-01-08 | 山东信通电子股份有限公司 | Target tracking method and device |
| CN112435276B (en)* | 2020-11-13 | 2024-08-23 | 鹏城实验室 | Vehicle tracking method and device, intelligent terminal and storage medium |
| CN112507949A (en)* | 2020-12-18 | 2021-03-16 | 北京百度网讯科技有限公司 | Target tracking method and device, road side equipment and cloud control platform |
| CN112528925B (en)* | 2020-12-21 | 2024-05-07 | 深圳云天励飞技术股份有限公司 | Pedestrian tracking and image matching method and related equipment |
| CN114913386B (en)* | 2021-01-29 | 2025-08-19 | 北京原创世代科技有限公司 | Training method of multi-target tracking model and multi-target tracking method |
| CN112837349A (en)* | 2021-02-09 | 2021-05-25 | 普联技术有限公司 | Target tracking method, target tracking equipment and computer-readable storage medium |
| CN113033447A (en)* | 2021-04-02 | 2021-06-25 | 蔚来汽车科技(安徽)有限公司 | Method for tracking object in video frame sequence, automatic parking method and device thereof |
| CN113792697B (en)* | 2021-09-23 | 2023-09-05 | 重庆紫光华山智安科技有限公司 | Target detection method, target detection device, electronic equipment and readable storage medium |
| CN114359341A (en)* | 2021-12-29 | 2022-04-15 | 湖南国科微电子股份有限公司 | Multi-target tracking method and device, terminal equipment and readable storage medium |
| CN114581491B (en)* | 2022-04-30 | 2022-07-22 | 苏州浪潮智能科技有限公司 | Pedestrian trajectory tracking method, system and related device |
| CN115080551B (en)* | 2022-06-15 | 2024-05-03 | 苏州轻棹科技有限公司 | Target track management method and device |
| CN115423846A (en)* | 2022-09-26 | 2022-12-02 | 青岛以萨数据技术有限公司 | Multi-target track tracking method and device |
| CN115965657B (en)* | 2023-02-28 | 2023-06-02 | 安徽蔚来智驾科技有限公司 | Target tracking method, electronic device, storage medium and vehicle |
- 2023
- 2023-02-28CNCN202310171954.8Apatent/CN115965657B/enactiveActive
- 2023-12-19WOPCT/CN2023/140010patent/WO2024179141A1/enactivePending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102065275A (en)* | 2009-11-17 | 2011-05-18 | 中国科学院电子学研究所 | Multi-target tracking method in intelligent video monitoring system |
| CN115311330A (en)* | 2019-10-11 | 2022-11-08 | 杭州云栖智慧视通科技有限公司 | Video multi-target tracking method based on position prediction |
| WO2021134172A1 (en)* | 2019-12-30 | 2021-07-08 | 华为技术有限公司 | Trajectory prediction method and related device |
| WO2022127180A1 (en)* | 2020-12-17 | 2022-06-23 | 深圳云天励飞技术股份有限公司 | Target tracking method and apparatus, and electronic device and storage medium |
| WO2022142417A1 (en)* | 2020-12-31 | 2022-07-07 | 深圳云天励飞技术股份有限公司 | Target tracking method and apparatus, electronic device, and storage medium |
| WO2022142918A1 (en)* | 2020-12-31 | 2022-07-07 | 深圳云天励飞技术股份有限公司 | Multi-target tracking system and method |
| CN114169241A (en)* | 2021-12-09 | 2022-03-11 | 北京邮电大学 | An end-to-end multi-target recognition, tracking and prediction method |
Non-Patent Citations (1)
| Title |
|---|
| 基于实时递推最小二乘的多目标编批研究;姜鑫;陈武雄;朱明;郝志成;高文;;国外电子测量技术(第02期);全文* |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115965657A (en) | 2023-04-14 |
| WO2024179141A1 (en) | 2024-09-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115965657B (en) | Target tracking method, electronic device, storage medium and vehicle | |
| CN108875666B (en) | Method and device for acquiring motion trail, computer equipment and storage medium | |
| CN106951484B (en) | Picture retrieval method and device, computer equipment and computer readable medium | |
| CN111079533B (en) | Unmanned vehicle driving decision method, unmanned vehicle driving decision device and unmanned vehicle | |
| CN112805661B (en) | Vehicle entry detection | |
| CN114612987A (en) | A method and device for facial expression recognition | |
| CN112634316B (en) | Target tracking method, device, equipment and storage medium | |
| CN112734931A (en) | Method and system for assisting point cloud target detection | |
| CN113657596B (en) | Method and device for training model and image recognition | |
| CN115620090A (en) | Model training method, low-illumination target re-identification method and device, and terminal equipment | |
| CN112650300A (en) | Unmanned aerial vehicle obstacle avoidance method and device | |
| CN112558036A (en) | Method and apparatus for outputting information | |
| CN117784798B (en) | Target tracking method, intelligent device and computer readable storage medium | |
| CN117574179B (en) | Method and device for constructing multi-task learning model | |
| CN115965944B (en) | Target information detection method, device, driving device and medium | |
| CN116861262B (en) | Perception model training method and device, electronic equipment and storage medium | |
| CN117991283A (en) | Multi-target tracking detection method, equipment and medium | |
| CN116702085A (en) | Target trajectory fusion method, device, electronic equipment, and storage medium | |
| CN111860070A (en) | Method and apparatus for identifying changed objects | |
| CN115792943A (en) | Positioning method, device, readable storage medium and vehicle for self-driving vehicle | |
| CN115713112A (en) | Multi-target tracking method for target domain data based on space topology and identity aggregation | |
| CN114626519A (en) | Model training method and device and image generation method | |
| CN117291952B (en) | Multi-target tracking method and device based on speed prediction and image reconstruction | |
| CN116558540B (en) | Model training method and device, and track generating method and device | |
| CN118690187A (en) | A method, device, terminal device and storage medium for predicting target state |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |