CN102254429B - Video identification-based detection method of detection device of violation vehicles - Google Patents

Abstract

Translated fromChinese

本发明公开了一种基于视频识别的违章车辆检测装置和方法。检测系统包括上位管理机、控制与传输装置、红外线全景摄像机和高速特写摄像机，其中控制与传输装置包括视频采集和转换模块、规则设置模块、车辆检测模块、违章分析模块、摄像机控制模块和数据无线传输模块。检测方法应用图像识别技术、视频追踪技术，有效地解决了对违章车辆的多目标检测问题，可以避免对正常变道行为的误判，并且可以对夜间和恶劣气象条件下较远距离的违章车辆进行检测。本发明所采用的设备和技术成熟，检测效率高，大大降低了检测成本，提高了交通管理的效率和智能化水平。

The invention discloses a device and method for detecting illegal vehicles based on video recognition. The detection system includes an upper management computer, a control and transmission device, an infrared panoramic camera and a high-speed close-up camera, wherein the control and transmission device includes a video acquisition and conversion module, a rule setting module, a vehicle detection module, a violation analysis module, a camera control module and a data wireless Transport module. The detection method uses image recognition technology and video tracking technology to effectively solve the problem of multi-target detection of illegal vehicles, avoid misjudgment of normal lane-changing behaviors, and detect long-distance illegal vehicles at night and in bad weather conditions. to test. The equipment and technology adopted by the invention are mature, the detection efficiency is high, the detection cost is greatly reduced, and the efficiency and intelligent level of traffic management are improved.

Description

Translated fromChinese

一种应用基于视频识别的违章车辆检测装置的违章车辆检测方法A method for detecting illegal vehicles using a video recognition-based illegal vehicle detection device

技术领域technical field

本发明属于交通违章自动检测技术领域，具体涉及一种基于视频识别的违章车辆检测装置和方法。 The invention belongs to the technical field of automatic detection of traffic violations, in particular to a device and method for detecting illegal vehicles based on video recognition. the

背景技术Background technique

随着我国城市汽车保有量的快速增加，交通违章行为也越来越频繁的发生。特别是对于道路交叉口，由于进口道车道数较多，而各车道车辆排队长度不均，恶意加塞和违章压实线变道等行为极为普遍，不仅影响了长排队车流的正常行驶，也易阻塞短排队车流所在车道，更影响了遵守规章驾车人的公平心理，极易造成车辆碰擦事故和驾驶员纠纷。2010年4月1日起，新修改施行的《机动车驾驶证申领和使用规定》中，对临时变道加塞被查处的驾驶人，将处以罚款200元，记2分。但各地交管部门在实际中的执行效果并不理想。因此，对城市道路交叉口车辆违章变道行为的监控有着重大的实际意义。 With the rapid increase of urban car ownership in our country, traffic violations are also occurring more and more frequently. Especially for road intersections, due to the large number of entrance lanes and the uneven queuing length of vehicles in each lane, behaviors such as malicious jamming and illegal compaction line changing are very common, which not only affects the normal driving of long-queuing traffic, but also easily Blocking the lane where the short-queue traffic flow is located affects the fairness of drivers who abide by the rules, and it is very easy to cause vehicle collision accidents and driver disputes. From April 1, 2010, in the newly revised and implemented "Regulations on Application and Use of Motor Vehicle Driving Licenses", drivers who are investigated for temporarily changing lanes and blocking traffic will be fined 200 yuan and scored 2 points. However, the implementation effect of traffic control departments in various places is not ideal. Therefore, monitoring the illegal lane-changing behavior of vehicles at urban road intersections has great practical significance. the

但是，目前尚没有成熟的交叉口进口道车辆加塞违章行为的自动化检测技术。 However, there is no mature automatic detection technology for vehicle jamming violations at intersections. the

我国当前对交叉口进口道违章行为的检测普遍采用的是感应线圈技术，即在交叉口停止线或中央分隔线等处埋设感应线圈，通过感应线圈触发照相拍摄。这种系统有以下几方面缺陷：一是需要开挖道路埋设线圈，如果出现故障或线圈本身使用寿命结束，需要开挖路面翻修或更换；二是由于交叉口进口道不可变道实线往往很长，距离停止线长达数十米，因此对违章变道控制来说，埋设线圈的成本非常高；三是在距离停止线较远的车道线为虚线的区域，由于无法区分恶意加塞和正常变道行为，若仍然采用感应线圈进行检测，易造成误拍并导致纠纷。 At present, the detection of violations at intersections and entrances in my country generally adopts the induction coil technology, that is, the induction coil is buried at the intersection stop line or the central dividing line, and the camera is triggered by the induction coil. This system has the following defects: one is that it is necessary to excavate the road to bury the coil, if there is a failure or the service life of the coil itself ends, the road surface needs to be excavated for renovation or replacement; Long, tens of meters away from the stop line, so for illegal lane change control, the cost of embedding coils is very high; third, in the area where the lane line far from the stop line is a dotted line, it is impossible to distinguish between malicious traffic jams and normal traffic jams. If the behavior of changing lanes is still detected by induction coils, it is easy to cause false shots and lead to disputes. the

另一种交叉口交通违章行为检测方法是通过视频技术对交通违章行为进行监测，即检测设备一端接入红绿灯信号，另一端接入摄像头，对闯红灯等违章行为进行拍照。但该类技术有几个缺陷：一是普遍缺少视频追踪技术和模式识别技术，只能根据红绿灯信号捕捉闯红灯行为，难以捕捉实线变道和恶意加塞行为，适用性低；二是对摄像机的拍摄范围精度要求高，如果摄像范围越过停止线，则误拍的几率会很高，且摄像机拍摄画面中的其他车辆易对后期处理带来干扰；三是摄像机视线易被遮挡，难以捕捉大型车后方违章的小型车。 Another method for detecting traffic violations at intersections is to monitor traffic violations through video technology, that is, one end of the detection device is connected to the traffic light signal, and the other end is connected to a camera to take pictures of violations such as running a red light. However, this type of technology has several shortcomings: first, it generally lacks video tracking technology and pattern recognition technology, and can only capture red light running behaviors based on traffic light signals, and it is difficult to capture solid line lane changes and malicious blocking behaviors, and its applicability is low; The accuracy of the shooting range is high. If the shooting range crosses the stop line, the probability of taking a wrong shot will be high, and other vehicles in the camera's shooting screen will easily interfere with the post-processing; the third is that the camera's line of sight is easily blocked, making it difficult to capture large vehicles. A small car with a rear violation. the

专利CN101593422B提出了一种视频追踪和模式识别的交通违章检测方法，该方法中采用的模式识别技术及算法尚不成熟，研发风险大，难以保证检测的正确率和准确率。该检测方法对多目标同时违章的情况更是无法应付。同时，该检测方法难以应对高峰期跟车较紧的车队，无法对牌照受前车遮挡的违章车辆进行拍照。该检测方法亦无法在恶劣天气下正常工作。 Patent CN101593422B proposes a traffic violation detection method based on video tracking and pattern recognition. The pattern recognition technology and algorithm used in this method are immature, and the research and development risks are high, so it is difficult to guarantee the accuracy and accuracy of detection. This detection method is especially unable to cope with the situation that multiple targets violate the rules at the same time. At the same time, this detection method is difficult to deal with the fleets that are closely following vehicles during peak hours, and cannot take pictures of illegal vehicles whose license plates are blocked by the vehicle in front. The detection method also does not work properly in bad weather. the

发明内容Contents of the invention

技术问题：本发明的目的是提供一种能准确检测多目标车辆同时出现违章实现变道、虚线实线恶意加塞等交通违章行为、并对违章车辆进行有效拍照的基于视频识别的违章车辆检测装置和方法。 Technical problem: The purpose of the present invention is to provide a video recognition-based illegal vehicle detection device that can accurately detect traffic violations such as traffic violations such as lane change, dotted and solid lines, etc., and can effectively take pictures of the illegal vehicles. and methods.

技术方案：基于视频识别的违章车辆检测装置，包括用于规则定义、指令下达和数据管理分析的上位管理机；用于视频信号处理、摄像机控制和数据传输的控制与传输装置；用于视频采集和全景拍摄的红外线全景摄像机；用于违章车辆照片拍摄与数据传输的高速特写摄像机；所述上位管理机和控制与传输装置通过无线网络连接，所述控制与传输装置通过数据线分别与红外线全景摄像机和高速特写摄像机连接。所述的控制与传输装置包括视频采集和转换模块、规则设置模块、车辆检测模块、违章分析模块、摄像机控制模块和数据无线传输模块。数据无线传输模块与上位管理机通过无线网络连接，数据无线传输模块将上位管理机发布的规则和指令传送给规则设置模块，将视频采集和转换模块获取并存入内存的违章车辆照片传输给上位管理机；规则设置模块将上位管理机发布的规则和指令传送给视频采集和转换模块、车辆检测模块和违章分析模块；视频采集和转换模块通过数据线分别与红外线全景摄像机和高速特写摄像机连接；车辆检测模块对视频采集和转换模块获取的帧位图进行处理，并将处理结果发送给违章分析模块；违章分析模块对车辆检测模块发送来的检测结果进行综合判断，根据判断结果形成指令并将指令发送给摄像机控制模块，摄像机控制模块通过数据线分别与红外线全景摄像机和高速特写摄像机连接。 Technical solution: Violation vehicle detection device based on video recognition, including upper management computer for rule definition, order issuance and data management analysis; control and transmission device for video signal processing, camera control and data transmission; for video acquisition and an infrared panoramic camera for panoramic shooting; a high-speed close-up camera for photographing and data transmission of illegal vehicles; The video camera and the high-speed close-up camera are connected. The control and transmission device includes a video collection and conversion module, a rule setting module, a vehicle detection module, a violation analysis module, a camera control module and a data wireless transmission module. The data wireless transmission module is connected to the upper management machine through a wireless network. The data wireless transmission module transmits the rules and instructions issued by the upper management machine to the rule setting module, and transmits the photos of illegal vehicles acquired and stored in the memory by the video acquisition and conversion module to the upper level. Management machine; the rule setting module transmits the rules and instructions issued by the upper management machine to the video acquisition and conversion module, vehicle detection module and violation analysis module; the video acquisition and conversion module is respectively connected to the infrared panoramic camera and the high-speed close-up camera through data lines; The vehicle detection module processes the frame bitmap obtained by the video acquisition and conversion module, and sends the processing result to the violation analysis module; the violation analysis module comprehensively judges the detection results sent by the vehicle detection module, forms instructions according to the judgment results and sends The instruction is sent to the camera control module, and the camera control module is respectively connected with the infrared panoramic camera and the high-speed close-up camera through data lines.

基于视频识别的违章车辆检测方法，包括以下步骤： A method for detecting illegal vehicles based on video recognition, comprising the following steps:

1）红外线全景摄像机将连续拍摄的视频数据传送至视频采集和转换模块；1) The infrared panoramic camera transmits the continuous shooting video data to the video acquisition and conversion module;

2）视频采集和转换模块将接收到的视频数据转换成帧位图并存入内存；2) The video capture and conversion module converts the received video data into a frame bitmap and stores it in memory;

3）车辆检测模块从内存中提取帧位图，对车辆进行识别，检测车辆运动轨迹并发送给违章分析模块，当发现车辆触碰车道线时，进入步骤4）；3) The vehicle detection module extracts the frame bitmap from the memory, identifies the vehicle, detects the vehicle trajectory and sends it to the violation analysis module, and when the vehicle is found to touch the lane line, enter step 4);

4）如车辆触碰道口车道线，违章分析模块直接判断为违章，并进入步骤5）；如车辆触碰道路车道线，摄像机控制模块控制红外线全景摄像机追踪该车辆并将连续的多幅帧位图发送给违章分析模块，违章分析模块将车辆运动轨迹与规则设置模块中定义的连续压线时间限值和最小变道时间间隔等违章规则进行对比分析，当该车辆连续触碰同一道路车道线时间超过连续压线时间限值或触碰道路车道线后在最小变道时间间隔内又触碰另一条道路车道线时，违章分析模块判断为违章，并进入步骤5），否则回到步骤3）； 4) If the vehicle touches the lane line of the crossing, the violation analysis module directly judges it as a violation, and proceeds to step 5); if the vehicle touches the lane line of the road, the camera control module controls the infrared panoramic camera to track the vehicle and captures multiple consecutive frames The map is sent to the violation analysis module, which compares the vehicle trajectory with the violation rules defined in the rule setting module, such as the continuous line pressing time limit and the minimum lane change time interval, when the vehicle continuously touches the same road lane line When the time exceeds the time limit for continuous line pressing or touches another road lane line within the minimum lane change time interval after touching the road lane line, the violation analysis module judges it as a violation and enters step 5), otherwise returns to step 3 );

5）摄像机控制模块控制红外线全景摄像机和特定位置的高速特写摄像机锁定追踪违章车辆，对违章车辆拍照并将照片及违章车辆信息发送给视频采集和转换模块；5) The camera control module controls the infrared panoramic camera and the high-speed close-up camera at a specific location to lock and track the illegal vehicles, take pictures of the illegal vehicles and send the photos and information of the illegal vehicles to the video acquisition and conversion module;

6）视频采集和转换模块将违章车辆照片和信息存入内存并通过数据无线传输模块传送至上位管理机。6) The video acquisition and conversion module stores the photos and information of illegal vehicles into the memory and transmits them to the upper management computer through the data wireless transmission module.

本发明的步骤3）中，违章分析模块（25）使用顺序区域生长法根据行驶车辆是否遮挡视频模拟车道线来检测车辆触碰车道线；步骤4）中，摄像机控制模块（26）使用序列蒙特卡洛滤波算法控制红外线全景摄像机（3）追踪车辆。步骤4）中的连续压线时间限值和最小变道时间间隔等违章规则由交通管理部门确定。 In step 3) of the present invention, the violation analysis module (25) uses the sequential region growing method to detect whether the vehicle touches the lane line according to whether the driving vehicle blocks the video simulation lane line; in step 4), the camera control module (26) uses the sequence Monte The Carlo filter algorithm controls the infrared panoramic camera (3) to track the vehicle. In step 4), the violation rules such as the time limit of continuous line pressing and the minimum time interval of lane change are determined by the traffic management department. the

本发明的步骤5）中，红外线全景摄像机和特定位置的高速特写摄像机可对违章车辆进行瞬时抓拍和延时补拍，如延时若干秒后补拍，或通过视频追踪功能等违章车辆行驶至路口时补拍，以确保获取违章车辆的牌照信息。 In step 5) of the present invention, the infrared panoramic camera and the high-speed close-up camera at a specific position can instantly capture and delay re-shooting of the violating vehicle, such as re-shooting after a delay of several seconds, or through the video tracking function and other illegal vehicles driving to Make up shots at intersections to ensure that license plate information of violating vehicles is obtained. the

本发明的步骤5）中，当有多台车辆同时违章时，违章分析模块通过摄像机控制模块同时控制不同的高速特写摄像机锁定追踪各个违章车辆进行拍照，可确保准确检测记录多目标车辆同时违章。 In step 5) of the present invention, when multiple vehicles violate the rules at the same time, the violation analysis module simultaneously controls different high-speed close-up cameras to lock and track each violation vehicle to take pictures through the camera control module, so as to ensure accurate detection and recording of simultaneous violations of multiple target vehicles. the

有益效果：本发明可以同时检测车辆恶意加塞、实线变道和连续变更两条以上车道的交通违章行为，避免对正常变道行为的误判；由于高速特写摄像机是从侧后方对违规车辆进行拍摄，并通过接受控制与传输装置的指令在最佳视角时进行拍摄，因此视角不会被违规车辆周围车辆阻挡，也避免了仅使用单视角摄像机时由于目标被阻挡而失去追踪的问题；本发明中每部高速特写摄像机只负责一定视角和距离内的车辆拍摄，故而可以确保同时获取多辆违规车辆的信息，并在不良的气候条件下提高照片的清晰度，可以对夜间和恶劣气象条件下较远距离的违章车辆进行检测，同时有效地解决了对违章车辆的多目标检测问题。本发明所采用的设备和技术成熟，检测效率高，大大降低了检测成本，提高了交通管理的效率和智能化水平。  Beneficial effects: the present invention can simultaneously detect traffic violations such as malicious traffic jamming, solid-line lane change, and continuous change of two or more lanes, avoiding misjudgment of normal lane-changing behaviors; Shooting, and shooting at the best angle of view by accepting the instructions of the control and transmission device, so the angle of view will not be blocked by vehicles around the violating vehicle, and the problem of losing track due to blocked targets when only using a single-angle camera is also avoided; In the invention, each high-speed close-up camera is only responsible for shooting vehicles within a certain angle of view and distance, so it can ensure that the information of multiple illegal vehicles can be obtained at the same time, and the clarity of the photos can be improved under bad weather conditions, which can be used for night and bad weather conditions. At the same time, it effectively solves the problem of multi-target detection of illegal vehicles. The equipment and technology adopted by the invention are mature, the detection efficiency is high, the detection cost is greatly reduced, and the efficiency and intelligent level of traffic management are improved.

 附图说明Description of drawings

图1 为本发明检测装置的结构框图；Fig. 1 is the structural block diagram of detection device of the present invention;

图2 为本发明检测现场平面示意图；Fig. 2 is a schematic diagram of the detection site of the present invention;

图3 为本发明检测装置的原理框图；Fig. 3 is the functional block diagram of detection device of the present invention;

图4为本发明检测方法的流程图，图中模拟实线为道口车道线，模拟虚线为道路车道线；Fig. 4 is the flow chart of detection method of the present invention, and among the figure, simulated solid line is crossing lane line, and simulated dotted line is road lane line;

图中有：上位管理机1，控制与传输装置2，红外线全景摄像机3，高速特写摄像机4，数据无线传输模块21，规则设置模块22，视频采集和转换模块23，车辆检测模块24，违章分析模块25，摄像机控制模块26，恶意加塞车辆5，连续变更车道6，压实线变道车辆7。In the figure, there are: upper management machine 1, control andtransmission device 2, infraredpanoramic camera 3, high-speed close-up camera 4, data wireless transmission module 21, rule setting module 22, video collection and conversion module 23, vehicle detection module 24, violation analysis Module 25, camera control module 26, malicious blocking vehicle 5, continuous lane change 6, compactionline changing vehicle 7.

具体实施方式Detailed ways

基于视频识别的违章车辆检测装置，包括用于规则定义、指令下达和数据管理分析的上位管理机1；用于视频信号处理、摄像机控制和数据传输的控制与传输装置2；用于视频采集和全景拍摄的红外线全景摄像机3；用于违章车辆照片拍摄与数据传输的高速特写摄像机4；所述上位管理机1和控制与传输装置2通过无线网络连接，所述控制与传输装置2通过数据线分别与红外线全景摄像机3和高速特写摄像机4连接。所述的控制与传输装置2包括视频采集和转换模块23、规则设置模块22、车辆检测模块24、违章分析模块25、摄像机控制模块26和数据无线传输模块21。数据无线传输模块21与上位管理机1通过无线网络连接，数据无线传输模块21将上位管理机1发布的规则和指令传送给规则设置模块22，将视频采集和转换模块23获取并存入内存的违章车辆照片传输给上位管理机1；规则设置模块22将上位管理机（1）发布的规则和指令传送给视频采集和转换模块23、车辆检测模块24和违章分析模块25；视频采集和转换模块23通过数据线分别与红外线全景摄像机3和高速特写摄像机4连接，视频采集和转换模块23将红外线全景摄像机3拍摄的连续视频数据转换成帧位图并存入内存，将红外线全景摄像机3和高速特写摄像机4拍摄的违章车辆照片存入内存；车辆检测模块24负责对视频采集和转换模块23获取的帧位图进行处理，并将处理结果发送给违章分析模块25；违章分析模块25对车辆检测模块24输入的检测结果进行综合判断，根据判断结果形成指令并将指令发送给摄像机控制模块26，摄像机控制模块26通过数据线分别与红外线全景摄像机3和高速特写摄像机4连接。 Violation vehicle detection device based on video recognition, including upper management machine 1 for rule definition, instruction issuance and data management analysis; control andtransmission device 2 for video signal processing, camera control and data transmission; for video acquisition and Infraredpanoramic camera 3 for panorama shooting; high-speed close-upcamera 4 for photographing and data transmission of illegal vehicles; the host management machine 1 and the control andtransmission device 2 are connected through a wireless network, and the control andtransmission device 2 is connected through a data line Connect with infraredpanoramic camera 3 and high-speed close-upcamera 4 respectively. The control andtransmission device 2 includes a video acquisition and conversion module 23 , a rule setting module 22 , a vehicle detection module 24 , a violation analysis module 25 , a camera control module 26 and a data wireless transmission module 21 . The data wireless transmission module 21 is connected with the upper management machine 1 through a wireless network, and the data wireless transmission module 21 transmits the rules and instructions issued by the upper management machine 1 to the rule setting module 22, and the video acquisition and conversion module 23 is acquired and stored in the internal memory. The photo of the violating vehicle is transmitted to the upper management machine 1; the rule setting module 22 transmits the rules and instructions issued by the upper management machine (1) to the video acquisition and conversion module 23, the vehicle detection module 24 and the violation analysis module 25; the video acquisition and conversion module 23 are respectively connected with the infraredpanoramic camera 3 and the high-speed close-upcamera 4 through the data line, and the video acquisition and conversion module 23 converts the continuous video data taken by the infraredpanoramic camera 3 into a frame bitmap and stores it in memory, and the infraredpanoramic camera 3 and the high-speed close-up camera The photo of the illegal vehicle taken by the close-upcamera 4 is stored in the internal memory; the vehicle detection module 24 is responsible for processing the frame bitmap obtained by the video acquisition and conversion module 23, and sends the processing result to the illegal analysis module 25; the illegal analysis module 25 detects the vehicle The detection result input by the module 24 is comprehensively judged, an instruction is formed according to the judgment result and the instruction is sent to the camera control module 26, and the camera control module 26 is respectively connected with the infraredpanoramic camera 3 and the high-speed close-upcamera 4 through data lines. the

基于视频识别的违章车辆检测方法，包括以下步骤： A method for detecting illegal vehicles based on video recognition, comprising the following steps:

1）红外线全景摄像机3将连续拍摄的视频数据传送至视频采集和转换模块23；1) The infraredpanoramic camera 3 transmits the video data captured continuously to the video acquisition and conversion module 23;

2）视频采集和转换模块23将接收到的视频数据转换成帧位图并存入内存；2) The video capture and conversion module 23 converts the received video data into a frame bitmap and stores it in memory;

3）车辆检测模块从内存中提取帧位图，，对车辆进行识别，检测车辆运动轨迹并发送给违章分析模块25，当发现车辆触碰车道线时，进入步骤4）；3) The vehicle detection module extracts the frame bitmap from the memory, identifies the vehicle, detects the vehicle trajectory and sends it to the violation analysis module 25, and when the vehicle is found to touch the lane line, enter step 4);

4）如车辆触碰道口车道线，违章分析模块25直接判断为违章，并进入步骤5）；如车辆触碰道路车道线，摄像机控制模块26控制红外线全景摄像机（3）追踪该车辆并将连续的多幅帧位图发送给违章分析模块25，违章分析模块25将车辆运动轨迹与规则设置模块22中定义的连续压线时间限值和最小变道时间间隔等违章规则进行对比分析，当该车辆连续触碰同一道路车道线时间超过连续压线时间限值或触碰道路车道线后在最小变道时间间隔内又触碰另一条道路车道线时，违章分析模块25判断为违章，并进入步骤5），否则回到步骤3）； 4) If the vehicle touches the lane line of the crossing, the violation analysis module 25 directly determines that it is a violation, and proceeds to step 5); if the vehicle touches the lane line of the road, the camera control module 26 controls the infrared panoramic camera (3) to track the vehicle and continuously A plurality of frame bitmaps are sent to the violation analysis module 25, and the violation analysis module 25 compares and analyzes the violation rules such as the continuous line-pressing time limit and the minimum lane change time interval defined in the vehicle trajectory and the rule setting module 22, when the When the vehicle continuously touches the lane line of the same road for more than the continuous line pressing time limit or touches another lane line within the minimum lane change time interval after touching the lane line, the violation analysis module 25 determines that it is a violation and enters the step 5), otherwise go back to step 3);

5）违章分析模块25通过摄像机控制模块26控制红外线全景摄像机3和特定位置的高速特写摄像机4锁定追踪违章车辆，对违章车辆进行瞬时抓拍和延时补拍并将照片及违章车辆信息存入内存；当有多台车辆同时违章时，违章分析模块25通过摄像机控制模块26同时控制不同的高速特写摄像机4对各个违章车辆进行拍照；5) The violation analysis module 25 controls the infraredpanoramic camera 3 and the high-speed close-upcamera 4 at a specific location to lock and track the violation vehicles through the camera control module 26, and takes instant snapshots and delayed supplementary shots of the violation vehicles, and stores the photos and information of the violation vehicles into the memory ; When multiple vehicles violated the rules simultaneously, the violation analysis module 25 controlled different high-speed close-upcameras 4 to take pictures of each violation vehicles by the camera control module 26 simultaneously;

6）违章分析模块25将违章车辆照片和信息通过数据无线传输模块21传送至上位管理机1。6) The violation analysis module 25 transmits the photos and information of the violation vehicle to the upper management machine 1 through the data wireless transmission module 21 .

本发明的违章车辆检测方法实施前，先要通过上位管理机1定义违章模式、道口车道线和道路车道线信息、监控区域等违章规则，并将违章规则通过数据无线传输模块传送至规则设置模块，规则设置模块将接收到的违章规则存储备用。 Before the violation vehicle detection method of the present invention is implemented, the violation rules such as the violation mode, crossing lane lines and road lane line information, and monitoring areas must be defined through the host management machine 1, and the violation rules are transmitted to the rule setting module through the data wireless transmission module , the rule setting module stores the received violation rules for backup. the

Claims (1)

Translated fromChinese

1.一种应用基于视频识别的违章车辆检测装置的违章车辆检测方法，其特征在于，该方法包括以下步骤：1. A method for detecting illegal vehicles using a video recognition-based vehicle illegal detection device, characterized in that the method may further comprise the steps:1）红外线全景摄像机（3）将连续拍摄的视频数据传送至视频采集和转换模块（23）；1) The infrared panoramic camera (3) transmits the continuous shooting video data to the video acquisition and conversion module (23);2）视频采集和转换模块（23）将接收到的视频数据转换成帧位图并存入内存；2) The video capture and conversion module (23) converts the received video data into a frame bitmap and stores it in memory;3）车辆检测模块从内存中提取帧位图，对车辆进行识别，检测车辆运动轨迹并发送给违章分析模块（25），当发现车辆触碰车道线时，进入步骤4）；3) The vehicle detection module extracts the frame bitmap from the memory, recognizes the vehicle, detects the vehicle trajectory and sends it to the violation analysis module (25), and when the vehicle is found to touch the lane line, enter step 4);4）如车辆触碰道口车道线，违章分析模块（25）直接判断为违章，并进入步骤5）；如车辆触碰道路车道线，摄像机控制模块（26）控制红外线全景摄像机（3）追踪该车辆并将连续的多幅帧位图发送给违章分析模块（25），违章分析模块（25）将车辆运动轨迹与规则设置模块（22）中定义的连续压线时间限值和最小变道时间间隔等违章规则进行对比分析，当该车辆连续触碰同一道路车道线时间超过连续压线时间限值或触碰道路车道线后在最小变道时间间隔内又触碰另一条道路车道线时，违章分析模块（25）判断为违章，并进入步骤5），否则回到步骤3）；4) If the vehicle touches the lane line of the crossing, the violation analysis module (25) directly judges it as a violation, and proceeds to step 5); if the vehicle touches the lane line of the road, the camera control module (26) controls the infrared panoramic camera (3) to track the The vehicle sends continuous multiple frame bitmaps to the violation analysis module (25), and the violation analysis module (25) compares the vehicle trajectory with the continuous line-marking time limit and the minimum lane change time defined in the rule setting module (22) When the vehicle continuously touches the lane line of the same road for more than the continuous line pressing time limit or touches the lane line of the road and then touches another lane line within the minimum lane change time interval, The violation analysis module (25) determines that it is a violation, and enters step 5), otherwise returns to step 3);5）摄像机控制模块（26）控制红外线全景摄像机（3）和特定位置的高速特写摄像机（4）锁定追踪违章车辆，对违章车辆拍照并将照片及违章车辆信息发送给视频采集和转换模块（23）；5) The camera control module (26) controls the infrared panoramic camera (3) and the high-speed close-up camera (4) at a specific location to lock and track the violating vehicle, take pictures of the violating vehicle and send the photo and the information of the violating vehicle to the video acquisition and conversion module (23 );6）视频采集和转换模块（23）将违章车辆照片和信息存入内存并通过数据无线传输模块（21）传送至上位管理机（1）。6) The video acquisition and conversion module (23) stores the photos and information of illegal vehicles into the memory and transmits them to the upper management machine (1) through the data wireless transmission module (21).所述步骤3）中，违章分析模块（25）使用顺序区域生长法根据行驶车辆是否遮挡视频模拟车道线来检测车辆触碰车道线；所述步骤4）中，摄像机控制模块（26）使用序列蒙特卡洛滤波算法控制红外线全景摄像机（3）追踪车辆。In the step 3), the violation analysis module (25) uses the sequential region growing method to detect whether the vehicle touches the lane line according to whether the driving vehicle blocks the video simulation lane line; in the step 4), the camera control module (26) uses the sequence The Monte Carlo filtering algorithm controls the infrared panoramic camera (3) to track the vehicle.

Images