CN117315632A - Non-motor vehicle dangerous driving behavior detection method and system based on time sequence data - Google Patents
Non-motor vehicle dangerous driving behavior detection method and system based on time sequence data Download PDFInfo
- Publication number
- CN117315632A CN117315632A CN202311202804.5A CN202311202804A CN117315632A CN 117315632 A CN117315632 A CN 117315632A CN 202311202804 A CN202311202804 A CN 202311202804A CN 117315632 A CN117315632 A CN 117315632A
- Authority
- CN
- China
- Prior art keywords
- motor vehicle
- video stream
- stream data
- dangerous driving
- lane
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/59—Context or environment of the image inside of a vehicle, e.g. relating to seat occupancy, driver state or inner lighting conditions
- G06V20/597—Recognising the driver's state or behaviour, e.g. attention or drowsiness
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
- G06V20/584—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads of vehicle lights or traffic lights
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Traffic Control Systems (AREA)
Abstract
本发明涉及驾驶行为检测，具体涉及基于时序数据的非机动车危险驾驶行为检测方法及系统，获取非机动车行驶的视频流数据；基于视频流数据获取目标场景中非机动车的轨迹信息，并根据轨迹信息对非机动车危险驾驶行为进行检测，得到第一危险驾驶行为检测结果；判断并获取视频流数据中非机动车由非机动车道行驶至机动车道或横穿道路的目标视频流数据；基于目标视频流数据获取多帧图像中非机动车的转向灯的空间信息，并基于空间信息获取多帧图像中非机动车的转向灯的时序信息，根据时序信息获取非机动车的转向灯状态；本发明提供的技术方案能够有效克服现有技术所存在的不能准确、高效地对非机动车危险驾驶行为进行检测的缺陷。
The invention relates to driving behavior detection, and specifically relates to a non-motor vehicle dangerous driving behavior detection method and system based on time series data, which obtains video stream data of non-motor vehicle driving; obtains the trajectory information of non-motor vehicles in a target scene based on the video stream data, and Detect the dangerous driving behavior of non-motor vehicles based on the trajectory information to obtain the first dangerous driving behavior detection result; determine and obtain the target video stream data of the non-motor vehicle driving from the non-motor vehicle lane to the motor vehicle lane or crossing the road in the video stream data; Obtain the spatial information of the non-motor vehicle's turn signal in the multi-frame image based on the target video stream data, obtain the temporal information of the non-motor vehicle's turn signal in the multi-frame image based on the spatial information, and obtain the non-motor vehicle's turn signal status based on the temporal information. ; The technical solution provided by the present invention can effectively overcome the shortcomings of the existing technology that cannot accurately and efficiently detect dangerous driving behaviors of non-motor vehicles.
Description
技术领域Technical field
本发明涉及驾驶行为检测，具体涉及基于时序数据的非机动车危险驾驶行为检测方法及系统。The invention relates to driving behavior detection, and in particular to a non-motor vehicle dangerous driving behavior detection method and system based on time series data.
背景技术Background technique
非机动车具有绿色环保、价格低廉、出行便捷等特点，近年来在人们的日常出行方式中占据着越来越高的比例。但是，由于非机动车出行群体交通安全意识淡薄等原因，导致在日常的出行中产生了大量交通违章行为，严重危害公众的人身和财产安全，给城市道路出行安全带来了巨大的挑战。Non-motorized vehicles are green, environmentally friendly, low-priced, and convenient for travel. In recent years, they have occupied an increasing proportion in people's daily travel methods. However, due to reasons such as the low traffic safety awareness of non-motor vehicle travel groups, a large number of traffic violations occur in daily travel, seriously endangering the personal and property safety of the public, and posing huge challenges to urban road travel safety.
当前在城市交通管理中，针对非机动车的危险驾驶行为缺乏有效的检测及管制措施。目前，对于非机动车的管理主要还是依赖于人工手段：通过交通管理人员人为地去现场进行监督管理，人力成本非常高，并且排查不精准，效率非常低。Currently, in urban traffic management, there is a lack of effective detection and control measures for dangerous driving behaviors of non-motor vehicles. At present, the management of non-motor vehicles mainly relies on manual means: traffic managers manually go to the site to supervise and manage, which has very high labor costs, inaccurate inspections, and very low efficiency.
发明内容Contents of the invention
(一)解决的技术问题(1) Technical problems solved
针对现有技术所存在的上述缺点，本发明提供了基于时序数据的非机动车危险驾驶行为检测方法及系统，能够有效克服现有技术所存在的不能准确、高效地对非机动车危险驾驶行为进行检测的缺陷。In view of the above shortcomings of the existing technology, the present invention provides a method and system for detecting dangerous driving behaviors of non-motor vehicles based on time series data, which can effectively overcome the inability of the existing technology to detect dangerous driving behaviors of non-motor vehicles accurately and efficiently. Detect defects.
(二)技术方案(2) Technical solutions
为实现以上目的，本发明通过以下技术方案予以实现：In order to achieve the above objectives, the present invention is achieved through the following technical solutions:
基于时序数据的非机动车危险驾驶行为检测方法，包括以下步骤：The non-motor vehicle dangerous driving behavior detection method based on time series data includes the following steps:
S1、获取非机动车行驶的视频流数据；S1. Obtain video stream data of non-motor vehicles;
S2、基于视频流数据获取目标场景中非机动车的轨迹信息，并根据轨迹信息对非机动车危险驾驶行为进行检测，得到第一危险驾驶行为检测结果；S2. Obtain the trajectory information of non-motor vehicles in the target scene based on the video stream data, and detect the dangerous driving behavior of non-motor vehicles based on the trajectory information to obtain the first dangerous driving behavior detection result;
S3、判断并获取视频流数据中非机动车由非机动车道行驶至机动车道或横穿道路的目标视频流数据；S3. Determine and obtain the target video stream data of the non-motor vehicle driving from the non-motor vehicle lane to the motor vehicle lane or crossing the road in the video stream data;
S4、基于目标视频流数据获取多帧图像中非机动车的转向灯的空间信息，并基于空间信息获取多帧图像中非机动车的转向灯的时序信息，根据时序信息获取非机动车的转向灯状态；S4. Obtain the spatial information of the non-motor vehicle's turn signal in the multi-frame image based on the target video stream data, obtain the temporal information of the non-motor vehicle's turn signal in the multi-frame image based on the spatial information, and obtain the non-motor vehicle's steering light based on the temporal information. light status;
S5、根据目标视频流数据及非机动车的转向灯状态，得到第二危险驾驶行为检测结果；S5. Obtain the second dangerous driving behavior detection result based on the target video stream data and the turn signal status of the non-motor vehicle;
S6、综合第一危险驾驶行为检测结果和第二危险驾驶行为检测结果，得到非机动车危险驾驶行为检测数据。S6. Combine the first dangerous driving behavior detection results and the second dangerous driving behavior detection results to obtain non-motor vehicle dangerous driving behavior detection data.
优选地，S2中基于视频流数据获取目标场景中非机动车的轨迹信息，并根据轨迹信息对非机动车危险驾驶行为进行检测，得到第一危险驾驶行为检测结果，包括：Preferably, in S2, the trajectory information of non-motor vehicles in the target scene is obtained based on the video stream data, and the dangerous driving behavior of non-motor vehicles is detected based on the trajectory information, and the first dangerous driving behavior detection result is obtained, including:
基于视频流数据获取目标场景中非机动车的缩略图数据，若未检测到头盔，则获取非机动车驾驶员的头部图像；Based on the video stream data, the thumbnail data of the non-motor vehicle in the target scene is obtained. If the helmet is not detected, the head image of the non-motor vehicle driver is obtained;
对头部图像进行灰度值分析，若灰度值分析结果满足第一预设阈值，则判断存在未佩戴头盔的危险驾驶行为。Perform gray value analysis on the head image. If the gray value analysis result meets the first preset threshold, it is determined that there is dangerous driving behavior without wearing a helmet.
优选地，S2中基于视频流数据获取目标场景中非机动车的轨迹信息，并根据轨迹信息对非机动车危险驾驶行为进行检测，得到第一危险驾驶行为检测结果，还包括：Preferably, S2 acquires the trajectory information of non-motor vehicles in the target scene based on the video stream data, detects the dangerous driving behavior of non-motor vehicles based on the trajectory information, and obtains the first dangerous driving behavior detection result, which also includes:
基于视频流数据获取目标场景中非机动车的跟踪框信息，基于跟踪框信息确定非机动车的行车轨迹点；Obtain the tracking frame information of non-motorized vehicles in the target scene based on video stream data, and determine the driving trajectory points of non-motorized vehicles based on the tracking frame information;
基于视频流数据获取目标场景中的车道停止线和信号灯状态，若在红灯状态下，行车轨迹点出现在车道停止线内，并跨过车道停止线出现在对面路口，则判断存在闯红灯的危险驾驶行为。Based on the video stream data, the lane stop line and signal light status in the target scene are obtained. If under the red light state, the driving trajectory point appears within the lane stop line and crosses the lane stop line and appears at the opposite intersection, it is judged that there is a danger of running the red light. driving behavior.
优选地，S2中基于视频流数据获取目标场景中非机动车的轨迹信息，并根据轨迹信息对非机动车危险驾驶行为进行检测，得到第一危险驾驶行为检测结果，还包括：Preferably, S2 acquires the trajectory information of non-motor vehicles in the target scene based on the video stream data, detects the dangerous driving behavior of non-motor vehicles based on the trajectory information, and obtains the first dangerous driving behavior detection result, which also includes:
基于视频流数据获取目标场景中非机动车的跟踪框信息，基于跟踪框信息确定非机动车的行车轨迹线；Obtain tracking frame information of non-motorized vehicles in the target scene based on video stream data, and determine the driving trajectory of non-motorized vehicles based on the tracking frame information;
基于视频流数据获取目标场景中的车道行驶方向，并计算行车轨迹线与车道行驶方向之间的夹角，若夹角大于第二预设阈值，则判断存在逆向行驶的危险驾驶行为。The lane driving direction in the target scene is obtained based on the video stream data, and the angle between the driving trajectory line and the lane driving direction is calculated. If the angle is greater than the second preset threshold, it is determined that there is dangerous driving behavior in the opposite direction.
优选地，S2中基于视频流数据获取目标场景中非机动车的轨迹信息，并根据轨迹信息对非机动车危险驾驶行为进行检测，得到第一危险驾驶行为检测结果，还包括：Preferably, S2 acquires the trajectory information of non-motor vehicles in the target scene based on the video stream data, detects the dangerous driving behavior of non-motor vehicles based on the trajectory information, and obtains the first dangerous driving behavior detection result, which also includes:
基于视频流数据获取目标场景中非机动车的跟踪框信息，基于跟踪框信息确定非机动车的行车轨迹点；Obtain the tracking frame information of non-motorized vehicles in the target scene based on video stream data, and determine the driving trajectory points of non-motorized vehicles based on the tracking frame information;
基于视频流数据获取目标场景中的机动车车道线，若有大于预设数量的行车轨迹点出现在同一机动车道内，则判断存在占用机动车道行驶的危险驾驶行为。The motor vehicle lane lines in the target scene are obtained based on the video stream data. If more than the preset number of driving track points appear in the same motor vehicle lane, it is judged that there is dangerous driving behavior occupying the motor vehicle lane.
优选地，S3中判断并获取视频流数据中非机动车由非机动车道行驶至机动车道或横穿道路的目标视频流数据，包括：Preferably, S3 determines and obtains the target video stream data of a non-motorized vehicle traveling from a non-motorized lane to a motorized lane or crossing a road in the video stream data, including:
对视频流数据进行检测，确定非机动车由非机动车道行驶至机动车道的变道时刻；Detect the video stream data to determine the lane change moment when a non-motorized vehicle travels from a non-motorized lane to a motorized lane;
将视频流数据中变道时刻前后预设时长的图像序列确定为目标视频流数据。The image sequence of a preset duration before and after the lane change moment in the video stream data is determined as the target video stream data.
优选地，所述对视频流数据进行检测，确定非机动车由非机动车道行驶至机动车道的变道时刻，包括：Preferably, the detection of video stream data to determine the lane change moment when a non-motor vehicle travels from a non-motor vehicle lane to a motor vehicle lane includes:
基于视频流数据获取目标场景中非机动车的车辆区域，以及各车道的车道线信息；Obtain the non-motorized vehicle area in the target scene and the lane line information of each lane based on the video stream data;
确定车辆区域与各车道之间的重叠区域，根据车辆区域与各车道之间的重叠区域的变化情况确定非机动车由非机动车道行驶至机动车道。Determine the overlapping area between the vehicle area and each lane, and determine the non-motorized vehicle traveling from the non-motorized lane to the motorized lane based on changes in the overlapping area between the vehicle area and each lane.
优选地，所述根据车辆区域与各车道之间的重叠区域的变化情况确定非机动车由非机动车道行驶至机动车道，包括：Preferably, the step of determining whether a non-motor vehicle travels from a non-motor vehicle lane to a motor vehicle lane based on changes in the overlap area between the vehicle area and each lane includes:
若车辆区域与非机动车道之间的重叠区域越来越小，同时车辆区域与非机动车道相邻机动车道之间的重叠区域越来越大，则判断存在变道行驶的危险驾驶行为；If the overlapping area between the vehicle area and the non-motorized lane is getting smaller and smaller, and at the same time, the overlapping area between the vehicle area and the motorized lane adjacent to the non-motorized lane is getting larger and larger, it is judged that there is dangerous driving behavior of changing lanes;
若预设时间段内，出现预设次数的变道行驶行为，则判断存在横穿道路的危险驾驶行为。If a preset number of lane-changing driving behaviors occur within a preset time period, it is judged that there is dangerous driving behavior across the road.
优选地，S4中基于目标视频流数据获取多帧图像中非机动车的转向灯的空间信息，并基于空间信息获取多帧图像中非机动车的转向灯的时序信息，根据时序信息获取非机动车的转向灯状态，包括：Preferably, in S4, the spatial information of the turn signal of non-motor vehicles in the multi-frame image is obtained based on the target video stream data, the temporal information of the turn signal of the non-motor vehicle in the multi-frame image is obtained based on the spatial information, and the non-motor vehicle turn signal is obtained based on the temporal information. The turn signal status of the motor vehicle includes:
将目标视频流数据输入训练好的检测网络，检测网络进行特征提取，输出多帧图像中非机动车的转向灯的空间信息；Input the target video stream data into the trained detection network, and the detection network performs feature extraction and outputs the spatial information of the turn signals of non-motor vehicles in multi-frame images;
基于空间信息对多帧图像中的相邻帧图像进行差分处理，得到差分序列，根据差分序列得到多帧图像中非机动车的转向灯的时序信息；Based on the spatial information, perform differential processing on adjacent frame images in the multi-frame images to obtain a differential sequence, and obtain the timing information of the turn signals of non-motor vehicles in the multi-frame images based on the differential sequence;
将时序信息输入训练好的检测网络，检测网络进行特征提取，输出非机动车由非机动车道行驶至机动车道或横穿道路过程中的转向灯状态。Input the time series information into the trained detection network, and the detection network performs feature extraction and outputs the turn signal status of the non-motorized vehicle when it travels from the non-motorized lane to the motorized lane or crosses the road.
基于时序数据的非机动车危险驾驶行为检测系统，包括视频流数据获取模块、第一危险驾驶行为检测模块、目标视频流数据获取模块、转向灯状态获取模块、第二危险驾驶行为检测模块和非机动车危险驾驶行为检测结果输出模块；The non-motor vehicle dangerous driving behavior detection system based on time series data includes a video stream data acquisition module, a first dangerous driving behavior detection module, a target video stream data acquisition module, a turn signal status acquisition module, a second dangerous driving behavior detection module and a non-motor vehicle dangerous driving behavior detection module. Motor vehicle dangerous driving behavior detection result output module;
视频流数据获取模块，获取非机动车行驶的视频流数据；The video stream data acquisition module obtains the video stream data of non-motor vehicle driving;
第一危险驾驶行为检测模块，基于视频流数据获取目标场景中非机动车的轨迹信息，并根据轨迹信息对非机动车危险驾驶行为进行检测，得到第一危险驾驶行为检测结果；The first dangerous driving behavior detection module obtains the trajectory information of non-motor vehicles in the target scene based on video stream data, and detects the dangerous driving behavior of non-motor vehicles based on the trajectory information to obtain the first dangerous driving behavior detection results;
目标视频流数据获取模块，判断并获取视频流数据中非机动车由非机动车道行驶至机动车道或横穿道路的目标视频流数据；The target video stream data acquisition module determines and obtains the target video stream data of the non-motor vehicle driving from the non-motor vehicle lane to the motor vehicle lane or crossing the road in the video stream data;
转向灯状态获取模块，基于目标视频流数据获取多帧图像中非机动车的转向灯的空间信息，并基于空间信息获取多帧图像中非机动车的转向灯的时序信息，根据时序信息获取非机动车的转向灯状态；The turn signal status acquisition module obtains the spatial information of the non-motor vehicle turn signals in the multi-frame images based on the target video stream data, and obtains the temporal information of the non-motor vehicle turn signals in the multi-frame images based on the spatial information, and obtains the non-motor vehicle turn signals based on the temporal information. Motor vehicle turn signal status;
第二危险驾驶行为检测模块，根据目标视频流数据及非机动车的转向灯状态，得到第二危险驾驶行为检测结果；The second dangerous driving behavior detection module obtains the second dangerous driving behavior detection result based on the target video stream data and the turn signal status of the non-motor vehicle;
非机动车危险驾驶行为检测结果输出模块，综合第一危险驾驶行为检测结果和第二危险驾驶行为检测结果，得到非机动车危险驾驶行为检测数据。The non-motor vehicle dangerous driving behavior detection result output module combines the first dangerous driving behavior detection result and the second dangerous driving behavior detection result to obtain non-motor vehicle dangerous driving behavior detection data.
(三)有益效果(3) Beneficial effects
与现有技术相比，本发明所提供的基于时序数据的非机动车危险驾驶行为检测方法及系统，具有以下有益效果：Compared with the existing technology, the non-motor vehicle dangerous driving behavior detection method and system based on time series data provided by the present invention have the following beneficial effects:
1)获取非机动车行驶的视频流数据，基于视频流数据获取目标场景中非机动车的轨迹信息，并根据轨迹信息对非机动车危险驾驶行为进行检测，得到第一危险驾驶行为检测结果，从而能够根据视频流数据中非机动车的轨迹信息，实现对未佩戴头盔、闯红灯、逆向行驶、占用机动车道行驶等非机动车危险驾驶行为进行准确、高效的检测；1) Obtain the video stream data of non-motor vehicles driving, obtain the trajectory information of non-motor vehicles in the target scene based on the video stream data, and detect the dangerous driving behavior of non-motor vehicles based on the trajectory information, and obtain the first dangerous driving behavior detection result, This enables accurate and efficient detection of dangerous driving behaviors of non-motor vehicles such as not wearing helmets, running red lights, driving in the opposite direction, and occupying motor vehicle lanes based on the trajectory information of non-motor vehicles in the video stream data;
2)判断并获取视频流数据中非机动车由非机动车道行驶至机动车道或横穿道路的目标视频流数据，基于目标视频流数据获取多帧图像中非机动车的转向灯的空间信息，并基于空间信息获取多帧图像中非机动车的转向灯的时序信息，根据时序信息获取非机动车的转向灯状态，根据目标视频流数据及非机动车的转向灯状态，得到第二危险驾驶行为检测结果，从而能够根据目标视频流数据及非机动车的转向灯状态，实现对变道行驶、横穿道路、不按照规定使用转向灯等非机动车危险驾驶行为进行准确、高效的检测。2) Determine and obtain the target video stream data of the non-motor vehicle driving from the non-motor vehicle lane to the motor vehicle lane or crossing the road in the video stream data, and obtain the spatial information of the turn signal of the non-motor vehicle in the multi-frame image based on the target video stream data. And based on the spatial information, the timing information of the non-motor vehicle's turn light in the multi-frame image is obtained, the non-motor vehicle's turn light status is obtained based on the timing information, and the second dangerous driving is obtained based on the target video stream data and the non-motor vehicle's turn light status. Behavior detection results can be used to accurately and efficiently detect dangerous driving behaviors of non-motor vehicles such as changing lanes, crossing roads, and not using turn signals in accordance with regulations based on the target video stream data and the turn signal status of non-motor vehicles.
附图说明Description of drawings
为了更清楚地说明本发明实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍。显而易见地，下面描述中的附图仅仅是本发明的一些实施例，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 more clearly explain the embodiments of the present invention or the technical solutions in the prior art, the drawings needed to describe the embodiments or the prior art will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without exerting creative efforts.
图1为本发明的流程示意图；Figure 1 is a schematic flow diagram of the present invention;
图2为本发明中根据目标视频流数据及非机动车的转向灯状态得到第二危险驾驶行为检测结果的流程示意图。Figure 2 is a schematic flowchart of obtaining the second dangerous driving behavior detection result according to the target video stream data and the turn signal status of the non-motor vehicle in the present invention.
具体实施方式Detailed ways
为使本发明实施例的目的、技术方案和优点更加清楚，下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述。显然，所描述的实施例是本发明一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。In order to make the purpose, technical solutions and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts fall within the scope of protection of the present invention.
基于时序数据的非机动车危险驾驶行为检测方法，如图1所示，①获取非机动车行驶的视频流数据。The non-motor vehicle dangerous driving behavior detection method based on time series data is shown in Figure 1. ① Obtain the video stream data of non-motor vehicle driving.
②基于视频流数据获取目标场景中非机动车的轨迹信息，并根据轨迹信息对非机动车危险驾驶行为进行检测，得到第一危险驾驶行为检测结果。② Obtain the trajectory information of non-motor vehicles in the target scene based on the video stream data, and detect the dangerous driving behavior of non-motor vehicles based on the trajectory information to obtain the first dangerous driving behavior detection result.
1)基于视频流数据获取目标场景中非机动车的轨迹信息，并根据轨迹信息对非机动车危险驾驶行为进行检测，得到第一危险驾驶行为检测结果，包括：1) Obtain the trajectory information of non-motor vehicles in the target scene based on video stream data, and detect the dangerous driving behavior of non-motor vehicles based on the trajectory information to obtain the first dangerous driving behavior detection results, including:
基于视频流数据获取目标场景中非机动车的缩略图数据，若未检测到头盔，则获取非机动车驾驶员的头部图像；Based on the video stream data, the thumbnail data of the non-motor vehicle in the target scene is obtained. If the helmet is not detected, the head image of the non-motor vehicle driver is obtained;
对头部图像进行灰度值分析，若灰度值分析结果满足第一预设阈值，则判断存在未佩戴头盔的危险驾驶行为。Perform gray value analysis on the head image. If the gray value analysis result meets the first preset threshold, it is determined that there is dangerous driving behavior without wearing a helmet.
2)基于视频流数据获取目标场景中非机动车的轨迹信息，并根据轨迹信息对非机动车危险驾驶行为进行检测，得到第一危险驾驶行为检测结果，还包括：2) Obtain the trajectory information of non-motor vehicles in the target scene based on the video stream data, and detect the dangerous driving behavior of non-motor vehicles based on the trajectory information, and obtain the first dangerous driving behavior detection result, which also includes:
基于视频流数据获取目标场景中非机动车的跟踪框信息，基于跟踪框信息确定非机动车的行车轨迹点；Obtain the tracking frame information of non-motorized vehicles in the target scene based on video stream data, and determine the driving trajectory points of non-motorized vehicles based on the tracking frame information;
基于视频流数据获取目标场景中的车道停止线和信号灯状态，若在红灯状态下，行车轨迹点出现在车道停止线内，并跨过车道停止线出现在对面路口，则判断存在闯红灯的危险驾驶行为。Based on the video stream data, the lane stop line and signal light status in the target scene are obtained. If under the red light state, the driving trajectory point appears within the lane stop line and crosses the lane stop line and appears at the opposite intersection, it is judged that there is a danger of running the red light. driving behavior.
3)基于视频流数据获取目标场景中非机动车的轨迹信息，并根据轨迹信息对非机动车危险驾驶行为进行检测，得到第一危险驾驶行为检测结果，还包括：3) Obtain the trajectory information of non-motor vehicles in the target scene based on the video stream data, and detect the dangerous driving behavior of non-motor vehicles based on the trajectory information, and obtain the first dangerous driving behavior detection result, which also includes:
基于视频流数据获取目标场景中非机动车的跟踪框信息，基于跟踪框信息确定非机动车的行车轨迹线；Obtain tracking frame information of non-motorized vehicles in the target scene based on video stream data, and determine the driving trajectory of non-motorized vehicles based on the tracking frame information;
基于视频流数据获取目标场景中的车道行驶方向，并计算行车轨迹线与车道行驶方向之间的夹角，若夹角大于第二预设阈值，则判断存在逆向行驶的危险驾驶行为。The lane driving direction in the target scene is obtained based on the video stream data, and the angle between the driving trajectory line and the lane driving direction is calculated. If the angle is greater than the second preset threshold, it is determined that there is dangerous driving behavior in the opposite direction.
4)基于视频流数据获取目标场景中非机动车的轨迹信息，并根据轨迹信息对非机动车危险驾驶行为进行检测，得到第一危险驾驶行为检测结果，还包括：4) Obtain the trajectory information of non-motor vehicles in the target scene based on the video stream data, and detect the dangerous driving behavior of non-motor vehicles based on the trajectory information, and obtain the first dangerous driving behavior detection result, which also includes:
基于视频流数据获取目标场景中非机动车的跟踪框信息，基于跟踪框信息确定非机动车的行车轨迹点；Obtain the tracking frame information of non-motorized vehicles in the target scene based on video stream data, and determine the driving trajectory points of non-motorized vehicles based on the tracking frame information;
基于视频流数据获取目标场景中的机动车车道线，若有大于预设数量的行车轨迹点出现在同一机动车道内，则判断存在占用机动车道行驶的危险驾驶行为。The motor vehicle lane lines in the target scene are obtained based on the video stream data. If more than the preset number of driving track points appear in the same motor vehicle lane, it is judged that there is dangerous driving behavior occupying the motor vehicle lane.
上述技术方案，获取非机动车行驶的视频流数据，基于视频流数据获取目标场景中非机动车的轨迹信息，并根据轨迹信息对非机动车危险驾驶行为进行检测，得到第一危险驾驶行为检测结果，从而能够根据视频流数据中非机动车的轨迹信息，实现对未佩戴头盔、闯红灯、逆向行驶、占用机动车道行驶等非机动车危险驾驶行为进行准确、高效的检测。The above technical solution obtains video stream data of non-motor vehicles driving, obtains trajectory information of non-motor vehicles in the target scene based on the video stream data, and detects dangerous driving behaviors of non-motor vehicles based on the trajectory information to obtain the first dangerous driving behavior detection As a result, it is possible to accurately and efficiently detect dangerous driving behaviors of non-motor vehicles such as not wearing helmets, running red lights, driving in the opposite direction, and occupying motor vehicle lanes based on the trajectory information of non-motor vehicles in the video stream data.
如图1和图2所示，③判断并获取视频流数据中非机动车由非机动车道行驶至机动车道或横穿道路的目标视频流数据，具体包括：As shown in Figures 1 and 2, ③ determine and obtain the target video stream data of non-motorized vehicles traveling from non-motorized lanes to motorized lanes or crossing roads in the video stream data, specifically including:
对视频流数据进行检测，确定非机动车由非机动车道行驶至机动车道的变道时刻；Detect the video stream data to determine the lane change moment when a non-motorized vehicle travels from a non-motorized lane to a motorized lane;
将视频流数据中变道时刻前后预设时长的图像序列确定为目标视频流数据。The image sequence of a preset duration before and after the lane change moment in the video stream data is determined as the target video stream data.
具体地，对视频流数据进行检测，确定非机动车由非机动车道行驶至机动车道的变道时刻，包括：Specifically, the video stream data is detected to determine the lane change moment when a non-motorized vehicle travels from a non-motorized lane to a motorized lane, including:
基于视频流数据获取目标场景中非机动车的车辆区域，以及各车道的车道线信息；Obtain the non-motorized vehicle area in the target scene and the lane line information of each lane based on the video stream data;
确定车辆区域与各车道之间的重叠区域，根据车辆区域与各车道之间的重叠区域的变化情况确定非机动车由非机动车道行驶至机动车道。Determine the overlapping area between the vehicle area and each lane, and determine the non-motorized vehicle traveling from the non-motorized lane to the motorized lane based on changes in the overlapping area between the vehicle area and each lane.
具体地，根据车辆区域与各车道之间的重叠区域的变化情况确定非机动车由非机动车道行驶至机动车道，包括：Specifically, determining whether a non-motor vehicle travels from a non-motor vehicle lane to a motor vehicle lane is based on changes in the overlap area between the vehicle area and each lane, including:
若车辆区域与非机动车道之间的重叠区域越来越小，同时车辆区域与非机动车道相邻机动车道之间的重叠区域越来越大，则判断存在变道行驶的危险驾驶行为；If the overlapping area between the vehicle area and the non-motorized lane is getting smaller and smaller, and at the same time, the overlapping area between the vehicle area and the motorized lane adjacent to the non-motorized lane is getting larger and larger, it is judged that there is dangerous driving behavior of changing lanes;
若预设时间段内，出现预设次数的变道行驶行为，则判断存在横穿道路的危险驾驶行为。If a preset number of lane-changing driving behaviors occur within a preset time period, it is judged that there is dangerous driving behavior across the road.
④基于目标视频流数据获取多帧图像中非机动车的转向灯的空间信息，并基于空间信息获取多帧图像中非机动车的转向灯的时序信息，根据时序信息获取非机动车的转向灯状态，具体包括：④Acquire the spatial information of the non-motor vehicle's turn signal in the multi-frame image based on the target video stream data, obtain the temporal information of the non-motor vehicle's turn signal in the multi-frame image based on the spatial information, and obtain the non-motor vehicle's turn signal based on the temporal information. status, specifically including:
将目标视频流数据输入训练好的检测网络，检测网络进行特征提取，输出多帧图像中非机动车的转向灯的空间信息；Input the target video stream data into the trained detection network, and the detection network performs feature extraction and outputs the spatial information of the turn signals of non-motor vehicles in multi-frame images;
基于空间信息对多帧图像中的相邻帧图像进行差分处理，得到差分序列，根据差分序列得到多帧图像中非机动车的转向灯的时序信息；Based on the spatial information, perform differential processing on adjacent frame images in the multi-frame images to obtain a differential sequence, and obtain the timing information of the turn signals of non-motor vehicles in the multi-frame images based on the differential sequence;
将时序信息输入训练好的检测网络，检测网络进行特征提取，输出非机动车由非机动车道行驶至机动车道或横穿道路过程中的转向灯状态。Input the time series information into the trained detection network, and the detection network performs feature extraction and outputs the turn signal status of the non-motorized vehicle when it travels from the non-motorized lane to the motorized lane or crosses the road.
⑤根据目标视频流数据及非机动车的转向灯状态，得到第二危险驾驶行为检测结果。⑤According to the target video stream data and the turn signal status of the non-motor vehicle, the second dangerous driving behavior detection result is obtained.
上述技术方案，判断并获取视频流数据中非机动车由非机动车道行驶至机动车道或横穿道路的目标视频流数据，基于目标视频流数据获取多帧图像中非机动车的转向灯的空间信息，并基于空间信息获取多帧图像中非机动车的转向灯的时序信息，根据时序信息获取非机动车的转向灯状态，根据目标视频流数据及非机动车的转向灯状态，得到第二危险驾驶行为检测结果，从而能够根据目标视频流数据及非机动车的转向灯状态，实现对变道行驶、横穿道路、不按照规定使用转向灯等非机动车危险驾驶行为进行准确、高效的检测。The above technical solution determines and obtains the target video stream data of a non-motor vehicle traveling from a non-motor vehicle lane to a motor vehicle lane or crossing the road in the video stream data, and obtains the space of the turn signal of the non-motor vehicle in the multi-frame image based on the target video stream data. information, and obtain the timing information of the turn signal of non-motor vehicles in multi-frame images based on spatial information, obtain the turn signal status of non-motor vehicles based on the timing information, and obtain the second Dangerous driving behavior detection results, so that based on the target video stream data and the turn signal status of non-motor vehicles, accurate and efficient detection of dangerous non-motor vehicle driving behaviors such as changing lanes, crossing roads, and not using turn signals in accordance with regulations can be achieved. detection.
⑥综合第一危险驾驶行为检测结果和第二危险驾驶行为检测结果，得到非机动车危险驾驶行为检测数据。⑥ Combine the first dangerous driving behavior detection results and the second dangerous driving behavior detection results to obtain non-motor vehicle dangerous driving behavior detection data.
本申请技术方案中，还公开了一种基于时序数据的非机动车危险驾驶行为检测系统，包括视频流数据获取模块、第一危险驾驶行为检测模块、目标视频流数据获取模块、转向灯状态获取模块、第二危险驾驶行为检测模块和非机动车危险驾驶行为检测结果输出模块；In the technical solution of this application, a non-motor vehicle dangerous driving behavior detection system based on time series data is also disclosed, including a video stream data acquisition module, a first dangerous driving behavior detection module, a target video stream data acquisition module, and a turn signal status acquisition module. module, a second dangerous driving behavior detection module and a non-motor vehicle dangerous driving behavior detection result output module;
视频流数据获取模块，获取非机动车行驶的视频流数据；The video stream data acquisition module obtains the video stream data of non-motor vehicle driving;
第一危险驾驶行为检测模块，基于视频流数据获取目标场景中非机动车的轨迹信息，并根据轨迹信息对非机动车危险驾驶行为进行检测，得到第一危险驾驶行为检测结果；The first dangerous driving behavior detection module obtains the trajectory information of non-motor vehicles in the target scene based on video stream data, and detects the dangerous driving behavior of non-motor vehicles based on the trajectory information to obtain the first dangerous driving behavior detection results;
目标视频流数据获取模块，判断并获取视频流数据中非机动车由非机动车道行驶至机动车道或横穿道路的目标视频流数据；The target video stream data acquisition module determines and obtains the target video stream data of the non-motor vehicle driving from the non-motor vehicle lane to the motor vehicle lane or crossing the road in the video stream data;
转向灯状态获取模块，基于目标视频流数据获取多帧图像中非机动车的转向灯的空间信息，并基于空间信息获取多帧图像中非机动车的转向灯的时序信息，根据时序信息获取非机动车的转向灯状态；The turn signal status acquisition module obtains the spatial information of the non-motor vehicle turn signals in the multi-frame images based on the target video stream data, and obtains the temporal information of the non-motor vehicle turn signals in the multi-frame images based on the spatial information, and obtains the non-motor vehicle turn signals based on the temporal information. Motor vehicle turn signal status;
第二危险驾驶行为检测模块，根据目标视频流数据及非机动车的转向灯状态，得到第二危险驾驶行为检测结果；The second dangerous driving behavior detection module obtains the second dangerous driving behavior detection result based on the target video stream data and the turn signal status of the non-motor vehicle;
非机动车危险驾驶行为检测结果输出模块，综合第一危险驾驶行为检测结果和第二危险驾驶行为检测结果，得到非机动车危险驾驶行为检测数据。The non-motor vehicle dangerous driving behavior detection result output module combines the first dangerous driving behavior detection result and the second dangerous driving behavior detection result to obtain non-motor vehicle dangerous driving behavior detection data.
以上实施例仅用以说明本发明的技术方案，而非对其限制；尽管参照前述实施例对本发明进行了详细的说明，本领域的普通技术人员应当理解：其依然可以对前述各实施例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同替换；而这些修改或者替换，并不会使相应技术方案的本质脱离本发明各实施例技术方案的精神和范围。The above embodiments are only used to illustrate the technical solutions of the present invention, but not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions of the foregoing embodiments. Modifications may be made to the recorded technical solutions, or equivalent substitutions may be made to some of the technical features; however, these modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of each embodiment of the present invention.
Claims (10)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311202804.5A CN117315632A (en) | 2023-09-18 | 2023-09-18 | Non-motor vehicle dangerous driving behavior detection method and system based on time sequence data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311202804.5A CN117315632A (en) | 2023-09-18 | 2023-09-18 | Non-motor vehicle dangerous driving behavior detection method and system based on time sequence data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117315632A true CN117315632A (en) | 2023-12-29 |
Family
ID=89285849
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311202804.5A Pending CN117315632A (en) | 2023-09-18 | 2023-09-18 | Non-motor vehicle dangerous driving behavior detection method and system based on time sequence data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117315632A (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113657295A (en) * | 2021-08-19 | 2021-11-16 | 上海商汤智能科技有限公司 | Vehicle behavior detection method, device and system |
| CN115187886A (en) * | 2022-03-16 | 2022-10-14 | 北京易华录信息技术股份有限公司 | Vehicle violation detection method and device and electronic equipment |
-
2023
- 2023-09-18 CN CN202311202804.5A patent/CN117315632A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113657295A (en) * | 2021-08-19 | 2021-11-16 | 上海商汤智能科技有限公司 | Vehicle behavior detection method, device and system |
| CN115187886A (en) * | 2022-03-16 | 2022-10-14 | 北京易华录信息技术股份有限公司 | Vehicle violation detection method and device and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109242251B (en) | Driving behavior safety detection method, device, equipment and storage medium | |
| CN101799987B (en) | Self-adaptive intelligent traffic light and control method thereof | |
| CN110619279B (en) | Road traffic sign instance segmentation method based on tracking | |
| CN114333330B (en) | Intersection event detection system based on road side edge holographic sensing | |
| CN103236191B (en) | Video-based safety precaution method for vehicle merging from highway ramp | |
| CN102622884B (en) | Vehicle illegal turning behavior detection method based on tracking | |
| Wang et al. | Investigating yielding behavior of heterogeneous vehicles at a semi-controlled crosswalk | |
| CN104376725B (en) | Signalized intersection non-motor vehicle lane channelizing control method under expansion effect | |
| CN110689724B (en) | Automatic motor vehicle zebra crossing present pedestrian auditing method based on deep learning | |
| CN202422425U (en) | Video-detection-based intelligent signal control system for crossing | |
| CN109299674B (en) | Tunnel illegal lane change detection method based on car lamp | |
| CN102496285A (en) | Method for determining red-light running of vehicles at intersection based on video detection and signal control system | |
| CN101236603A (en) | Vehicle Intelligent Rear View Method | |
| CN105005771A (en) | Method for detecting full line of lane based on optical flow point locus statistics | |
| CN103886755B (en) | Crossing exception parking rapid alarm system and method with the camera function that makes a dash across the red light | |
| CN104361648B (en) | Event data recorder provided with signal lamps for reminding other vehicles and control method of event data recorder | |
| CN111126171A (en) | Vehicle reverse running detection method and system | |
| CN102324183A (en) | Vehicle detection and grasp shoot method based on compound virtual coil | |
| CN108639059A (en) | Driver based on least action principle manipulates behavior quantization method and device | |
| CN113033275B (en) | Vehicle lane-changing non-turn signal lamp analysis system based on deep learning | |
| CN104915912B (en) | A kind of amber light dilemma zone modeling method | |
| CN104361649B (en) | Automobile data recorder with signal lamp for prompting other vehicles and control method thereof | |
| CN112818935B (en) | Multi-lane congestion detection and duration prediction method and system based on deep learning | |
| CN113283309A (en) | Method and system for identifying road construction state according to real-time pavement image | |
| CN104680618B (en) | A kind of drive recorder with the signal lighties for warning other vehicles and control method |
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 |