WO2021184570A1 - 电动自行车驾驶行为的识别方法、装置和计算机设备 - Google Patents

电动自行车驾驶行为的识别方法、装置和计算机设备 Download PDF

Info

Publication number
WO2021184570A1
WO2021184570A1 PCT/CN2020/098383 CN2020098383W WO2021184570A1 WO 2021184570 A1 WO2021184570 A1 WO 2021184570A1 CN 2020098383 W CN2020098383 W CN 2020098383W WO 2021184570 A1 WO2021184570 A1 WO 2021184570A1
Authority
WO
WIPO (PCT)
Prior art keywords
driving
electric bicycle
target electric
target
vehicle
Prior art date
Application number
PCT/CN2020/098383
Other languages
English (en)
French (fr)
Inventor
李云清
黄丹薇
Original Assignee
平安国际智慧城市科技股份有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 平安国际智慧城市科技股份有限公司 filed Critical 平安国际智慧城市科技股份有限公司
Publication of WO2021184570A1 publication Critical patent/WO2021184570A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/50—Context or environment of the image
    • G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
    • G06V20/54—Surveillance or monitoring of activities, e.g. for recognising suspicious objects of traffic, e.g. cars on the road, trains or boats
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/40—Scenes; Scene-specific elements in video content

Definitions

  • This application relates to the field of artificial intelligence technology, in particular to the field of computer vision technology, and in particular to a method, device, computer equipment and storage medium for identifying the driving behavior of an electric bicycle.
  • the identification of illegal behaviors of electric bicycles is generally through image recognition of multiple electric bicycle vehicle pictures captured by surveillance cameras installed in monitored areas, such as intersections, roads, etc., to determine whether there are illegal behaviors of electric bicycles. .
  • the inventor realized that if there are many vehicles in the captured picture, or the captured picture is affected by environmental factors such as weather and the picture is not clear, it is easy to cause misrecognition or miss recognition of the electric bicycle driving behavior, and then As a result, the recognition accuracy of electric bicycle driving behavior is low.
  • a method, a device, a computer device, and a storage medium for recognizing driving behavior of an electric bicycle are provided.
  • a method for identifying the driving behavior of an electric bicycle includes:
  • the dual-base recognition device is used to monitor the driving condition of the electric bicycle in the monitored area;
  • the vehicle position of the target electric bicycle is determined from the electronic map corresponding to the monitored area, and the first driving trajectory of the target electric bicycle through the monitored area is determined according to the vehicle position ;
  • the first driving trajectory and the second driving trajectory determine the target driving trajectory of the target electric bicycle through the monitored area, and determine the target electric bicycle to pass the monitored area according to the target driving trajectory Driving characteristics;
  • a recognition result of the driving behavior of the target electric bicycle is determined according to the driving characteristics.
  • An electric bicycle driving behavior identification device includes:
  • the information receiving module is used to receive the driving video and vehicle positioning information of the target electric bicycle driving through the monitored area sent by the dual-base recognition device; the dual-base recognition device is used to monitor the driving of the electric bicycle in the monitored area Condition;
  • the first driving trajectory determination module is used to determine the vehicle position of the target electric bicycle from the electronic map corresponding to the monitored area according to the driving video, and determine the target electric bicycle to pass through according to the vehicle position State the first vehicle track in the monitored area;
  • the second driving trajectory determination module is used to determine the positioning position of the electric bicycle in the electronic map according to the vehicle positioning information, and to determine the second driving of the electric bicycle passing the monitored area according to the positioning position Trajectory
  • a driving characteristic determination module configured to determine a target driving trajectory of the target electric bicycle passing through the monitored area according to the first driving trajectory and the second driving trajectory, and determining the electric bicycle according to the target driving trajectory The characteristics of driving through the monitored area;
  • the recognition result determination module is used to determine the recognition result of the driving behavior of the target electric bicycle according to the driving characteristics.
  • a computer device including a memory and one or more processors, the memory stores computer readable instructions, and when the computer readable instructions are executed by the processor, the one or more processors execute The following steps:
  • the dual-base recognition device is used to monitor the driving condition of the electric bicycle in the monitored area;
  • the vehicle position of the target electric bicycle is determined from the electronic map corresponding to the monitored area, and the first driving trajectory of the target electric bicycle through the monitored area is determined according to the vehicle position ;
  • the first driving trajectory and the second driving trajectory determine the target driving trajectory of the target electric bicycle through the monitored area, and determine the target electric bicycle to pass the monitored area according to the target driving trajectory Driving characteristics;
  • a recognition result of the driving behavior of the target electric bicycle is determined according to the driving characteristics.
  • One or more computer-readable storage media storing computer-readable instructions.
  • the one or more processors perform the following steps:
  • the dual-base recognition device is used to monitor the driving condition of the electric bicycle in the monitored area;
  • the vehicle position of the target electric bicycle is determined from the electronic map corresponding to the monitored area, and the first driving trajectory of the target electric bicycle through the monitored area is determined according to the vehicle position ;
  • the first driving trajectory and the second driving trajectory determine the target driving trajectory of the target electric bicycle through the monitored area, and determine the target electric bicycle to pass the monitored area according to the target driving trajectory Driving characteristics;
  • a recognition result of the driving behavior of the target electric bicycle is determined according to the driving characteristics.
  • the above method, device, computer equipment and storage medium for identifying electric bicycle driving behavior comprehensively consider the vehicle positioning information of the electric bicycle and the vehicle video, so that the recognized driving behavior of the electric bicycle is more accurate, and avoiding the comparison based only on the photographed pictures.
  • the recognition of electric bicycle driving behavior is prone to misrecognition or missing recognition, which leads to the defect of low recognition accuracy of electric bicycle driving behavior, and further improves the recognition accuracy of electric bicycle driving behavior.
  • Fig. 1 is an application scenario diagram of a method for identifying driving behavior of an electric bicycle according to one or more embodiments
  • FIG. 2 is a schematic flowchart of a method for recognizing driving behavior of an electric bicycle according to one or more embodiments
  • Figure 3 is a schematic diagram of a crossroad in accordance with one or more embodiments.
  • Fig. 4 is a schematic diagram of a preset passing area of a cross intersection according to one or more embodiments
  • FIG. 5 is a schematic flowchart of a method for recognizing driving behavior of an electric bicycle in another embodiment
  • Fig. 6 is a block diagram of a device for recognizing driving behavior of an electric bicycle according to one or more embodiments
  • Figure 7 is a block diagram of a computer device according to one or more embodiments.
  • the method for identifying the driving behavior of an electric bicycle can be applied to the application environment as shown in FIG. 1.
  • the dual-base identification device 110 and the server 120 communicate through a network.
  • the dual-base identification device 110 is set in the monitored area, such as each intersection of a cross road, for real-time monitoring of the driving conditions of the electric bicycle passing through the monitored area, for example, the dual-base identification device 110 collects the target electric bicycle driving through the monitored area
  • the driving video and vehicle positioning information, and the collected driving video and vehicle positioning information of the target electric bicycle traveling through the monitored area are sent to the corresponding server 120.
  • the server 120 determines the vehicle position of the target electric bicycle from the electronic map corresponding to the monitored area according to the driving video, and determines the first trajectory of the target electric bicycle through the monitored area according to the vehicle position; determines the target electric bicycle's location according to the vehicle positioning information According to the positioning position in the electronic map, the second driving trajectory of the target electric bicycle passing the monitored area is determined according to the positioning position; the target driving trajectory of the target electric bicycle passing the monitored area is determined according to the first driving trajectory and the second driving trajectory, according to the target The driving trajectory determines the driving characteristics of the target electric bicycle passing through the monitored area; the recognition result of the driving behavior of the target electric bicycle is determined according to the driving characteristics.
  • the dual-base identification device 110 refers to a radio frequency identification device installed with a camera device (such as a camera), such as a dual-base RFID (Radio Frequency Identification, radio frequency identification) device;
  • the server 120 may be an independent server or composed of multiple servers Server cluster to achieve.
  • a method for recognizing driving behavior of an electric bicycle is provided. Taking the method applied to the server in FIG. 1 as an example, the method includes the following steps:
  • Step S201 receiving the driving video and vehicle positioning information of the target electric bicycle traveling through the monitored area sent by the dual-base recognition device; the dual-base recognition device is used to monitor the driving condition of the electric bicycle in the monitored area.
  • the dual-base identification device refers to an RFID device with a camera, such as a dual-base RFID device; the dual-base identification device can monitor the driving conditions of electric bicycles in the monitored area, and collect the electric bicycles that have traveled through the monitored area. Bicycle vehicle information, driving video, vehicle positioning information, etc.; generally set at intersections, such as crossroads, T-shaped intersections, etc. As shown in Figure 3, one dual-base identification device is set up in the four directions of the intersection, and the antenna orientation of each dual-base identification device is clarified.
  • the dual-base identification device can also be installed in other locations, such as beside the road; the specific installation location depends on the monitored area, which is not specifically limited in this application.
  • Vehicle information generally refers to the license plate information, model information, owner information, etc. of electric bicycles.
  • Driving video refers to the video of electric bicycles traveling through the monitored area captured by the camera in the dual-base recognition device.
  • Vehicle positioning information refers to the The location information of electric bicycles that have traveled through the monitored area collected by the RFID device in the dual-base identification device may refer to the longitude and latitude coordinate information of the electric bicycle (such as longitude coordinate information and latitude coordinate information), such as (31.2121751783, 121.4411213954); It should be noted that the vehicle positioning information in this application refers to multiple vehicle positioning information.
  • an electric bicycle refers to a bicycle that uses a battery as an auxiliary energy source, which is equipped with an RFID tag, and the vehicle information, annual inspection information, insurance information, etc. of the electric bicycle can be written into the RFID tag through the RFID tag writing device;
  • the target electric bicycle refers to the electric bicycle that needs to recognize the driving behavior;
  • the monitored area refers to the monitoring area of the dual-base recognition device, such as intersections, roads, etc.
  • the dual-base recognition equipment set near the monitored area collects real-time driving video and vehicle positioning information of electric bicycles that have driven through the monitored area.
  • Both vehicle video and vehicle positioning information carry vehicle information; from the collected driving From the driving video and vehicle positioning information of electric bicycles passing through the monitored area, the driving video and vehicle positioning information belonging to the same vehicle information are screened out as the driving video and vehicle positioning information of the target electric bicycle driving through the monitored area.
  • the driving video and vehicle positioning information of the target electric bicycle that has traveled through the monitored area are sent to the corresponding server; the server receives the driving video and vehicle positioning information of the target electric bicycle that has traveled through the monitored area, so as to facilitate subsequent driving according to the monitored area
  • the driving video of the target electric bicycle and vehicle positioning information determine the driving behavior of the target electric bicycle.
  • an electric bicycle with an RFID tag installed passes through an intersection where a dual-base RFID device is installed.
  • the dual-base RFID device emits UHF electromagnetic wave signals, and the RFID tag installed on the electric bicycle receives UHF electromagnetic wave signals.
  • Start chip verification to write vehicle information, such as license plate information, and send the vehicle information back to the dual-base RFID device.
  • the dual-base RFID device communicates with the RFID tag installed on the electric bicycle multiple times to obtain multiple vehicle positioning information of the electric bicycle passing through the intersection; or, through RFID positioning technology, it can be passed Multiple vehicle positioning information of the electric bicycle at the intersection, each vehicle positioning information carries vehicle information, such as license plate information.
  • the dual-base RFID device can also start the camera to shoot the driving video of the electric bicycle passing the above-mentioned intersection, and filter the key frame static image of the vehicle from the driving video, extract the image characteristics of the key frame static image of the vehicle, and analyze the image characteristics , Obtain the vehicle information of the electric bicycle passing the above-mentioned intersection, such as license plate information; wherein, the static image of the vehicle key frame refers to a static image that can clearly display the license plate number of the electric bicycle.
  • the dual-base RFID equipment screens out the driving video of electric bicycles belonging to the same license plate information and multiple vehicle positioning information, as the target electric bicycle passing through the intersection
  • the driving video and vehicle positioning information of the target electric bicycle passing the intersection are sent to the corresponding server.
  • Step S202 Determine the vehicle position of the target electric bicycle from the electronic map corresponding to the monitored area according to the driving video, and determine the first trajectory of the target electric bicycle through the monitored area according to the vehicle position.
  • the electronic map refers to the map that maps the actual scene to the computer; the electronic map corresponding to the monitored area refers to the map that maps the actual scene of the monitored area to the computer, as shown in Figure 3
  • the electronic map of the intersection; the electronic map corresponding to the monitored area is beneficial to the subsequent determination of the target electric bicycle's trajectory.
  • the first driving trajectory refers to the driving trajectory determined based on the vehicle position of the target electric bicycle in the electronic map corresponding to the monitored area, which can reflect the driving behavior of the electric bicycle driving through the monitored area; for example, go straight, turn left, and turn left. Turn right and wait, as shown in Figure 3.
  • the server screens out the vehicle key frame static image of the target electric bicycle from the driving video, extracts the image characteristics of the vehicle key frame static image, analyzes the image characteristics, and determines the target electric bicycle in the image shown in the monitored area.
  • Location information according to the preset mapping relationship between the location information in the image shown in the monitored area and the location in the electronic map of the monitored area, determine the vehicle position of the target electric bicycle in the electronic map of the monitored area; compare each vehicle The position is connected as a position node, and the first trajectory of the target electric bicycle through the monitored area is obtained.
  • the camera of the dual-base recognition device since the camera of the dual-base recognition device has a fixed shooting angle and covers the entire monitored area, the size of the image shown in the monitored area corresponding to the video frame in the driving video is the same.
  • the server selects the static image of the vehicle key frame of the target electric bicycle from the driving video, inputs the static image of the vehicle key frame into the pre-trained convolutional neural network model, and extracts the key frame static image of the vehicle through the convolutional neural network model. Image features, and identify the image features to obtain the position information of the target electric bicycle in the monitored area.
  • the mapping relationship of the position in the electronic map is to determine the vehicle position of the target electric bicycle in the electronic map corresponding to the monitored area.
  • the vehicle position in the electronic map, and so on, can get multiple vehicle positions of the target electric bicycle in the electronic map corresponding to the monitored area; obtain the time stamp of the static image of the vehicle key frame as the time corresponding to the determined vehicle position Information: According to time information, connect each vehicle position as a position node to get the first trajectory of the electric bicycle through the monitored area.
  • Step S203 Determine the positioning position of the target electric bicycle in the electronic map according to the vehicle positioning information, and determine the second driving track of the target electric bicycle passing through the monitored area according to the positioning position.
  • the second driving trajectory refers to the driving trajectory determined based on the location of the target electric bicycle in the electronic map of the monitored area, which can reflect the driving behavior of the target electric bicycle through the monitored area; Turn left, turn right, etc., as shown in Figure 3.
  • the server removes duplicate vehicle positioning information in advance, and determines the positioning position of the target electric bicycle in the electronic map of the monitored area according to the mapping relationship between the preset positioning information and the positioning position in the electronic map of the intersection;
  • the positioning position is connected as a positioning node, and the second driving trajectory of the target electric bicycle through the monitored area is obtained.
  • the server queries the mapping relationship between the preset positioning information and the positioning position in the electronic map of the monitored area according to the vehicle positioning information, and determines the positioning position of the target electric bicycle in the electronic map corresponding to the monitored area; for example, the target electric bicycle
  • the vehicle positioning information is position A3, then according to the mapping relationship between the preset positioning information and the positioning position in the electronic map of the monitored area, the positioning position mapped to position A3 is determined to be position A4, and position A4 is used as the target electric bicycle
  • Step S204 Determine the target driving trajectory of the target electric bicycle passing the monitored area according to the first driving trajectory and the second driving trajectory, and determine the driving characteristics of the target electric bicycle passing the monitored area according to the target driving trajectory.
  • the target driving trajectory refers to the final driving trajectory comprehensively determined based on the first driving trajectory and the second driving trajectory, which can reflect the driving behavior of the target electric bicycle through the monitored area; for example, go straight, turn left, and turn right. Turn and so on, as shown in Figure 3.
  • the driving characteristics of the target electric bicycle passing through the monitored area are used to measure the driving information of the target electric bicycle passing through the monitored area, which can refer to the driving direction, driving area or traffic state, etc.; wherein the driving direction refers to passing straight through the intersection, or going backwards through the intersection , Turn left through intersections, etc.; driving area refers to main roads, restricted areas, secondary roads, motor vehicle lanes, etc.; traffic state refers to the state of the target electric bicycle passing through the intersection corresponding to the driving direction, such as running red lights, green lights, etc.; in actual In the scene, the driving characteristics refer to the characteristics of straight green light traffic, left turn green light traffic characteristics, straight red light characteristics, left turn red lights characteristics, retrograde passing characteristics, overloaded load characteristics, non-wearing helmet characteristics, upper bridge tunnel characteristics, and main road occupation characteristics , Driving into the prohibited and restricted area features, etc.
  • the server corrects the first driving trajectory according to the second driving trajectory, and obtains the corrected first driving trajectory as the target driving trajectory of the target electric bicycle passing through the monitored area; for example, the second half of the trajectory of the second driving trajectory , Connect with the first half of the trajectory of the first driving trajectory to obtain the corrected first driving trajectory as the target driving trajectory of the target electric bicycle passing through the monitored area; or, the server extracts the target electric bicycle from the first driving trajectory.
  • the monitored area includes the first vehicle position of each preset passing area, and the second vehicle position of the target electric bicycle in each preset passing area is extracted from the second driving track; acquiring the first vehicle of each preset passing area The position average value between the position and the second vehicle position is used as the vehicle position of each preset passing area; the vehicle position of each preset passing area is connected as a position node to obtain the target driving of the target electric bicycle through the monitored area Trajectory.
  • the target driving trajectory the driving direction, driving area and traffic state of the electric bicycle are determined as the driving characteristics of the target electric bicycle passing through the monitored area. In this way, comprehensively considering the first driving trajectory and the second driving trajectory of the target electric bicycle, the subsequent identification of the driving behavior of the target electric bicycle is more accurate, thereby improving the recognition accuracy of the driving behavior of the electric bicycle.
  • the server determines that the electric bicycle is going straight through the intersection and the current traffic light status is red according to the first driving trajectory of the target electric bicycle; according to the second driving trajectory of the target electric bicycle, it is determined that the electric bicycle is also going straight Passing an intersection; then confirm that the target driving trajectory of the target electric bicycle passing the intersection is going straight through a red light, and the driving characteristics are going straight and running a red light; for example, the server determines that the target electric bicycle is turning left and passing through the intersection according to the first driving trajectory of the target electric bicycle.
  • the current traffic light status is red; according to the second driving trajectory of the target electric bicycle, it is determined that the target electric bicycle is also turning left and passing the intersection, then confirm that the target driving trajectory of the target electric bicycle passing the intersection is turning left and running through a red light, and the driving characteristic is left Turn around and run the red light.
  • the server determines that the driving direction of the target electric bicycle is straight, and the driving area is the restricted area, which means that the driving characteristic of the target electric bicycle passing through the monitored area is straight through the restricted area. Restricted area.
  • the server can also use the human body structure, vehicle structure, and face structure algorithms to identify whether the target electric bicycle has overload behavior (such as overloading people, overloading objects, etc.) or not wearing it according to the panoramic large image in the driving video Helmet; at the same time, determine the license plate information and owner information of the target electric bicycle based on the vehicle information collected by the dual-base RFID device; if the target electric bicycle has overload behavior or does not wear a helmet, confirm that the target electric bicycle passes the driving characteristics of the monitored area as the target The electric bicycle is overloaded or the owner of the target electric bicycle is not wearing a helmet.
  • overload behavior such as overloading people, overloading objects, etc.
  • the server can also use the human body structure, vehicle structure, and face structure algorithms to identify whether the target electric bicycle has overload behavior (such as overloading people, overloading objects, etc.) or not wearing it according to the panoramic large image in the driving video Helmet; at the same time, determine the license plate information and owner information of the target electric bicycle based on the vehicle information collected by the dual-base
  • Step S205 Determine the recognition result of the driving behavior of the target electric bicycle according to the driving characteristics.
  • the server matches the driving characteristics with the preset illegal driving characteristics (such as turning left and running through the red light, running straight through the red light, etc.); if the matching is successful, it will confirm that the driving behavior of the target electric bicycle through the monitored area is the preset illegal driving Behavior; if the match fails, it is confirmed that the driving behavior of the target electric bicycle through the monitored area does not belong to the preset illegal driving behavior, indicating that the driving behavior of the target electric bicycle is legal driving behavior during the process of passing the monitored area, and the result is The recognition result of the driving behavior of the target electric bicycle.
  • the preset illegal driving characteristics such as turning left and running through the red light, running straight through the red light, etc.
  • the recognition result of the driving behavior of the target electric bicycle is determined, avoiding the recognition of the driving behavior of the electric bicycle based only on the captured pictures, which is prone to misrecognition or missed recognition, resulting in the electric bicycle
  • the defect of low recognition accuracy of driving behavior further improves the recognition accuracy of electric bicycle driving behavior; at the same time, the whole process does not require the participation of executive personnel, thus realizing automatic recognition of the driving behavior of electric bicycles passing through the monitored area
  • the purpose of this is to save a lot of labor costs, thereby improving the efficiency of identifying the driving behavior of the electric bicycle passing through the monitored area.
  • the purpose of automatically recognizing the driving behavior of the electric bicycle according to the vehicle positioning information and driving video of the electric bicycle is realized, combined with the first driving trajectory determined based on the driving video and the vehicle positioning information Determine the second driving trajectory, determine the target driving trajectory of the electric bicycle, and then determine the recognition result of the driving behavior of the electric bicycle, avoiding the recognition of the driving behavior of the electric bicycle based on the captured pictures, which is prone to misrecognition or missing recognition , Leading to the defect of low recognition accuracy of electric bicycle driving behavior, and further improving the recognition accuracy of electric bicycle driving behavior.
  • comprehensive consideration of the vehicle positioning information of the electric bicycle and the vehicle video makes the recognized driving behavior of the electric bicycle more accurate, thereby improving the recognition accuracy of the driving behavior of the electric bicycle.
  • the vehicle positioning information includes the corresponding time information. Then, in step S203, the positioning position of the target electric bicycle in the electronic map is determined according to the vehicle positioning information, and the first position of the target electric bicycle passing the monitored area is determined according to the positioning position. 2.
  • the driving trajectory includes: obtaining the location identifier of each vehicle's location information in the electronic map; according to the location identifier, determining the location of the target electric bicycle in the electronic map; according to the time information, each location location is connected as a positioning node , Get the second trajectory of the target electric bicycle through the monitored area.
  • the location identifier is used to identify the location of the vehicle positioning information on the electronic map of the monitored area.
  • the server obtains the position identifier of each vehicle positioning information in the electronic map of the monitored area; according to the position identifier, determines the position of each vehicle positioning information in the electronic map of the monitored area, as the target electric bicycle in the electronic map According to the time information corresponding to the vehicle positioning information, the correspondingly determined positioning positions are connected as positioning nodes in turn, and a line segment connecting each positioning node is obtained as the target electric bicycle passing through the second line of the monitored area Trajectory.
  • the second driving trajectory of the target electric bicycle passing through the monitored area is determined according to the vehicle positioning information, which is beneficial to subsequently combining the first driving trajectory of the electric bicycle passing the monitored area to automatically recognize the driving behavior of the electric bicycle It avoids the recognition of electric bicycle driving behavior based only on the captured pictures, which is prone to misrecognition or missed recognition, which leads to the defect of low recognition accuracy of electric bicycle driving behavior, and further improves the recognition accuracy of electric bicycle driving behavior.
  • the above step S204, according to the first driving trajectory and the second driving trajectory, determining the target driving trajectory of the target electric bicycle passing through the monitored area includes: extracting the target electric bicycle from the first driving trajectory when the target electric bicycle is being
  • the monitoring area includes the first vehicle position in the preset passing area; the second vehicle position of the target electric bicycle in the preset passing area is extracted from the second driving track; the target electric bicycle is determined according to the first vehicle position and the second vehicle position
  • the vehicle position of the bicycle in the preset passage area; the vehicle position of each preset passage area is connected as a position node to obtain the target driving trajectory of the target electric bicycle through the monitored area.
  • the preset passing area included in the monitored area refers to multiple areas divided by the monitored area, specifically referring to each passing area in the monitored area, such as dividing the entire intersection into multiple
  • the preset passage area is the small gray square area as shown in Figure 4.
  • the server identifies the preset pass area in the monitored area that the first vehicle trajectory passes, such as preset pass area A, preset pass area B, preset pass area C, etc.; extract the target from the first vehicle trajectory
  • the first vehicle position of the electric bicycle in the preset passing area for example, extracting the target electric bicycle from the first driving trajectory in the preset passing area A, the preset passing area B, and the preset passing area C as a vehicle position, respectively
  • the first vehicle position in the passage area A is (a, b), and the second vehicle
  • the target driving trajectory of the target electric bicycle passing through the intersection is determined, which is beneficial to subsequently accurately determining the driving behavior of the target electric bicycle passing through the monitored area according to the target driving trajectory. Improve the recognition accuracy of electric bicycle driving behavior.
  • the above step S204, determining the driving characteristics of the electric bicycle passing through the monitored area according to the target driving trajectory includes: determining the driving direction and driving area of the target electric bicycle according to the target driving trajectory; and obtaining the corresponding driving direction According to the current traffic light status of the intersection, the traffic status of the target electric bicycle is determined according to the current traffic light status; the driving direction, driving area and traffic status of the target electric bicycle are identified as the driving characteristics of the target electric bicycle passing the monitored area.
  • the server analyzes the target driving trajectory to obtain the driving direction and driving area of the electric bicycle, and obtains the current traffic light status of the intersection corresponding to the driving direction from the traffic light management system associated with the monitored area, such as corresponding to a straight intersection
  • the current traffic light status determines the traffic status of the target electric bicycle. For example, if the current traffic light is red, it means the traffic status of the target electric bicycle is running a red light; if the current traffic light is green, it means the target electric bicycle
  • the traffic state of the target electric bicycle is green light traffic; the determined driving direction, driving area and traffic state of the target electric bicycle are recognized as the driving characteristics of the target electric bicycle passing the monitored area.
  • the above step S205, determining the recognition result of the driving behavior of the target electric bicycle according to the driving characteristics includes: if the driving characteristics match the preset illegal driving characteristics, confirming that the driving behavior of the electric bicycle is illegal driving Behavior: If the driving characteristics do not match the preset illegal driving characteristics, the driving behavior of the electric bicycle is confirmed as a legal driving behavior.
  • the driving characteristics are used to characterize the driving information of the target electric bicycle passing through the monitored area, such as straight-going green light passing characteristics, left-turning green light passing characteristics, straight-going red light characteristics, left-turning red light characteristics, overloading characteristics, and not wearing a helmet Characteristics, etc.;
  • the preset illegal driving characteristics refer to the driving characteristics extracted from the preset illegal driving behavior information, such as the characteristics of driving straight through a red light, turning left and driving through a red light, and characteristics of overloading.
  • the server matches the driving characteristics with the preset illegal driving characteristics, and if the matching is successful, confirms that the driving characteristics of the target electric bicycle passing through the monitored area are the preset illegal driving characteristics, indicating that the driving behavior of the target electric bicycle is illegal Driving behavior; if the matching fails, it is confirmed that the driving characteristic of the target electric bicycle passing through the monitored area is not a preset illegal driving characteristic, indicating that the driving behavior of the target electric bicycle is a legal driving behavior.
  • the driving area is the main road
  • the target electric bicycle has illegal driving behavior of occupying the main road
  • the driving direction is straight and the traffic state is running a red light
  • the target electric bicycle has illegal driving behavior of driving straight and running a red light.
  • the entire process does not require the participation of executives, thereby achieving the purpose of automatically identifying the driving behavior of the target electric bicycle passing through the monitored area, thereby saving a lot of labor costs, and further improving the driving of the target electric bicycle passing through the intersection. Recognition efficiency of behavior.
  • the driving video carries vehicle information of the target electric bicycle; then, after confirming that the driving behavior of the target electric bicycle is an illegal driving behavior, it further includes: querying the preset vehicle according to the vehicle information of the target electric bicycle The corresponding relationship between the information and the owner's information is to obtain the owner's information of the target electric bicycle; obtain the preset reminder information; and send the preset reminder information to the terminal bound to the owner's information.
  • the owner information refers to information used to identify the identity of the owner, such as the owner's name, the owner's ID card number, etc.; the terminal bound to the owner's information may be a mobile phone, a tablet, etc.
  • the server queries the corresponding relationship between the preset license plate information and the owner information according to the license plate information of the target electric bicycle to obtain the owner information of the target electric bicycle; according to the illegal driving behavior of the target electric bicycle, Obtain the corresponding preset reminder information from the local database, such as "Please don't run the red light and drive safely"; send the preset reminder information to the owner's mobile phone corresponding to the owner's information in the form of SMS to remind the owner to drive safely.
  • the method further includes: uploading the illegal driving behavior of the target electric bicycle to the illegal driving behavior; the illegal driving platform is used for illegal driving according to the target electric bicycle Behavior, perform the corresponding violation processing operations.
  • the server can also upload the illegal driving behavior of the target electric bicycle to the illegal driving behavior; the illegal driving platform automatically executes the corresponding illegal driving behavior according to the illegal driving behavior of the target electric bicycle.
  • Illegal handling operations such as fines, deductions, etc.
  • the illegal handling platform can also store and record the illegal driving behavior of the target electric bicycle for backup and traceability.
  • the method for recognizing the driving behavior of the electric bicycle of the present application further includes: determining the driving behavior score of the target electric bicycle according to the recognition result of the driving behavior of the target electric bicycle; The corresponding relationship of the behavioral safety level determines the driving behavioral safety level of the target electric bicycle.
  • the driving behavior score and driving behavior safety level of the target electric bicycle are both used to measure the safety degree of the driving behavior of the driver corresponding to the target electric bicycle. If the driving behavior score is low, the driving behavior of the corresponding driver The safety level is low, indicating that the driver's driving behavior is dangerous.
  • the server recognizes the driving behavior of the target electric bicycle to determine the driving behavior of the target electric bicycle; query the correspondence between the preset driving behavior and the driving behavior score to obtain the driving behavior score of the target electric bicycle;
  • the driving behavior score is greater than or equal to the first threshold
  • the driving behavior safety level of the target electric bicycle is determined to be the first safety level
  • the driving behavior score of the target electric bicycle is greater than the second threshold and less than the first threshold
  • the driving of the target electric bicycle is determined
  • the behavioral safety level is the second safety level
  • the driving behavior safety level of the target electric bicycle is determined to be the third safety level.
  • the server may also send the driving behavior safety level of the target electric bicycle within a preset time range (such as one quarter, one year, etc.) to the car insurance server; the car insurance server is based on The driving behavior safety level of the target electric bicycle in the preset time range is determined, and the comprehensive driving behavior level of the target electric bicycle is determined; for example, the average value of the driving behavior safety level of the target electric bicycle in the preset time range is obtained as the target electric bicycle Comprehensive level of driving behavior; obtain the car insurance quotation information corresponding to the comprehensive level of driving behavior, and send the car insurance quotation information to the terminal bound to the owner's information for the owner to view.
  • a preset time range such as one quarter, one year, etc.
  • the auto insurance quote for the next year can be lowered; if the comprehensive level of driving behavior is low, it indicates that the owner’s daily driving behavior is not very good, you can Increase the auto insurance quotes for the next year.
  • the RFID sensing monitoring alarm device detects that the target electric bicycle enters an area that does not meet the electric bicycle charging, such as residential community lobbies, corridors, overhead floors, etc., then trigger Alarm, such as broadcasting alarm information, turning on alarm lights, etc.; and send the alarm information to the staff in the above-mentioned areas to remind the staff that there are currently electric bicycles that are not eligible for electric bicycle charging. Please go to the scene in time to deal with it.
  • Alarm such as broadcasting alarm information, turning on alarm lights, etc.
  • FIG. 5 another method for recognizing driving behavior of an electric bicycle is provided. Taking the method applied to the server in FIG. 1 as an example for description, the method includes the following steps:
  • Step S501 receiving the driving video and vehicle positioning information of the target electric bicycle traveling through the monitored area sent by the dual-base recognition device; the dual-base recognition device is used to monitor the driving condition of the electric bicycle in the monitored area.
  • Step S502 Determine the vehicle position of the target electric bicycle from the electronic map corresponding to the monitored area according to the driving video, and determine the first trajectory of the target electric bicycle through the monitored area according to the vehicle position.
  • Step S503 Determine the positioning position of the target electric bicycle in the electronic map according to the vehicle positioning information, and determine the second driving track of the target electric bicycle passing through the monitored area according to the positioning position.
  • Step S504 Determine the target driving trajectory of the target electric bicycle passing through the monitored area according to the first driving trajectory and the second driving trajectory, and determine the driving characteristics of the target electric bicycle passing through the monitored area according to the target driving trajectory.
  • step S505 if the driving characteristics of the vehicle match the preset illegal driving characteristics, it is confirmed that the driving behavior of the target electric bicycle is an illegal driving behavior.
  • Step S506 according to the vehicle information of the target electric bicycle, query the correspondence between the preset vehicle information and the owner information to obtain the owner information of the target electric bicycle; obtain the preset reminder information; send the preset reminder information to the owner information binding Fixed terminal.
  • Step S507 Upload the illegal driving behavior of the target electric bicycle to the illegal handling platform; the illegal handling platform is used to execute the corresponding illegal driving operation according to the illegal driving behavior of the target electric bicycle.
  • Step S508 Determine the driving behavior score of the target electric bicycle according to the recognition result of the driving behavior of the target electric bicycle; determine the driving behavior safety level of the target electric bicycle according to the corresponding relationship between the preset driving behavior score and the driving behavior safety level.
  • the above-mentioned electric bicycle driving behavior recognition method comprehensively considers the electric bicycle vehicle positioning information and the vehicle video, so that the recognized electric bicycle driving behavior is more accurate, and it is easy to recognize the electric bicycle driving behavior based on the captured pictures. There is a defect of misrecognition or missing recognition, leading to a low recognition accuracy rate of electric bicycle driving behavior, and further improving the recognition accuracy rate of electric bicycle driving behavior.
  • a device for identifying driving behavior of an electric bicycle including: an information receiving module 610, a first driving trajectory determining module 620, a second driving trajectory determining module 630, and driving characteristics
  • the determining module 640 and the recognition result determining module 650 wherein:
  • the information receiving module 610 is used to receive the driving video of the target electric bicycle driving through the monitored area and vehicle positioning information sent by the dual-base recognition device; the dual-base recognition device is used to monitor the driving condition of the electric bicycle passing through the monitored area.
  • the first driving trajectory determination module 620 is used to determine the vehicle position of the target electric bicycle from the electronic map corresponding to the monitored area according to the driving video, and determine the first driving trajectory of the target electric bicycle through the monitored area according to the vehicle position.
  • the second driving track determination module 630 is configured to determine the location of the target electric bicycle in the electronic map according to the vehicle positioning information, and determine the second driving track of the target electric bicycle through the monitored area according to the positioning position.
  • the driving characteristic determination module 640 is configured to determine the target driving trajectory of the target electric bicycle passing the monitored area according to the first driving trajectory and the second driving trajectory, and determine the driving characteristic of the target electric bicycle passing the monitored area according to the target driving trajectory.
  • the recognition result determination module 650 is configured to determine the recognition result of the driving behavior of the target electric bicycle according to the driving characteristics.
  • the vehicle positioning information includes corresponding time information; the second driving track determination module 630 is also used to obtain the position identifier of each vehicle positioning information in the electronic map; according to the position identifier, determine the location of the target electric bicycle The location location in the electronic map; according to time information, each location location is connected as a location node, and the second driving track of the target electric bicycle passing through the monitored area is obtained.
  • the driving characteristic determination module 640 is further configured to extract the first vehicle position of the target electric bicycle in the preset passing area included in the monitored area from the first driving trajectory; extract from the second driving trajectory The second vehicle position of the target electric bicycle in the preset passage area; the vehicle position of the target electric bicycle in the preset passage area is determined according to the first vehicle position and the second vehicle position; the vehicle position of each preset passage area is regarded as the position
  • the nodes are connected to obtain the target driving trajectory of the target electric bicycle passing the intersection; according to the target driving trajectory, the driving direction and driving area of the target electric bicycle are determined; the current traffic light state of the intersection corresponding to the driving direction is obtained, and according to the current traffic light state, Determine the traffic state of the target electric bicycle; identify the driving direction, driving area and traffic state of the target electric bicycle as the driving characteristics of the target electric bicycle passing through the monitored area.
  • the recognition result determination module 650 is further configured to confirm that the driving behavior of the target electric bicycle is an illegal driving behavior if the driving characteristics match the preset illegal driving characteristics; if the driving characteristics match the preset illegal driving characteristics If there is no match, it is confirmed that the driving behavior of the target electric bicycle is a legal driving behavior.
  • the driving video carries vehicle information of the target electric bicycle;
  • the electric bicycle driving behavior recognition device further includes an information sending module for querying preset vehicle information and owner information according to the vehicle information of the target electric bicycle The corresponding relationship between, obtain the owner information of the target electric bicycle; obtain the preset reminder information; send the preset reminder information to the terminal bound to the owner's information.
  • the device for identifying electric bicycle driving behavior further includes an upload module for uploading the illegal driving behavior of the target electric bicycle to the illegal driving behavior; the illegal driving platform is used to execute the illegal driving behavior of the target electric bicycle. Corresponding violation handling operations.
  • the electric bicycle driving behavior recognition device further includes a level determining module, which is used to determine the driving behavior score of the target electric bicycle according to the recognition result of the driving behavior of the target electric bicycle; and according to the preset driving behavior score The corresponding relationship with the safety level of driving behavior determines the safety level of driving behavior of the target electric bicycle.
  • Each of the above embodiments comprehensively considers the vehicle positioning information of the electric bicycle and the vehicle video, so that the recognized driving behavior of the electric bicycle is more accurate, avoiding the recognition of the driving behavior of the electric bicycle based only on the captured pictures, which is prone to misrecognition or Missing recognition leads to the defect that the recognition accuracy of electric bicycle driving behavior is low.
  • Each module in the above-mentioned electric bicycle driving behavior recognition device can be implemented in whole or in part by software, hardware and a combination thereof.
  • the above-mentioned modules may be embedded in the form of hardware or independent of the processor in the computer equipment, or may be stored in the memory of the computer equipment in the form of software, so that the processor can call and execute the operations corresponding to the above-mentioned modules.
  • a computer device is provided.
  • the computer device may be a server, and its internal structure diagram may be as shown in FIG. 7.
  • the computer equipment includes a processor, a memory, a network interface, and a database connected through a system bus. Among them, the processor of the computer device is used to provide calculation and control capabilities.
  • the memory of the computer device includes a non-volatile or volatile storage medium and internal memory.
  • the non-volatile or volatile storage medium stores an operating system, computer readable instructions, and a database.
  • the internal memory provides an environment for the operation of the operating system and computer-readable instructions in the non-volatile or volatile storage medium.
  • the database of the computer equipment is used to store data such as the driving video of the target electric bicycle, vehicle positioning information, driving trajectory, driving characteristics, and driving behavior recognition results.
  • the network interface of the computer device is used to communicate with an external terminal through a network connection.
  • FIG. 7 is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied.
  • the specific computer device may Including more or fewer parts than shown in the figure, or combining some parts, or having a different arrangement of parts.
  • a computer device includes a memory and one or more processors, in which computer-readable instructions are stored, and when the computer-readable instructions are executed by the processor, the method for identifying driving behavior of an electric bicycle provided in any one of the embodiments of the present application is realized A step of.
  • One or more computer-readable storage media storing computer-readable instructions.
  • the computer-readable storage media may be non-volatile or volatile.
  • the computer-readable instructions are executed by one or more processors , Enabling one or more processors to implement the steps of the method for identifying the driving behavior of an electric bicycle provided in any one of the embodiments of the present application.
  • Non-volatile memory may include read-only memory (Read-Only Memory, ROM), magnetic tape, floppy disk, flash memory, or optical storage.
  • Volatile memory may include random access memory (RAM) or external cache memory.
  • RAM can be in various forms, such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), etc.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Theoretical Computer Science (AREA)
  • Traffic Control Systems (AREA)

Abstract

本申请涉及一种电动自行车驾驶行为的识别方法，涉及人工智能技术领域，尤其涉及计算机视觉技术领域，包括：接收双基识别设备发送的目标电动自行车行驶过被监控区域的行车视频以及车辆定位信息；根据行车视频，从被监控区域对应的电子地图中确定出目标电动自行车的车辆位置，根据车辆位置确定目标电动自行车的第一行车轨迹；根据车辆定位信息确定目标电动自行车在电子地图中的定位位置，根据定位位置确定目标电动自行车的第二行车轨迹；根据第一行车轨迹和第二行车轨迹，确定目标行车轨迹，根据目标行车轨迹确定行车特征；根据行车特征确定对目标电动自行车的驾驶行为的识别结果。

Description

电动自行车驾驶行为的识别方法、装置和计算机设备
本申请要求于2020年03月16日提交中国专利局，申请号为202010181183.7，申请名称为“电动自行车驾驶行为的识别方法、装置和计算机设备”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及人工智能技术领域，尤其涉及计算机视觉技术领域，特别是涉及一种电动自行车驾驶行为的识别方法、装置、计算机设备和存储介质。
背景技术
随着电动自行车的普及，越来越多的人选择使用电动自行车作为出行的交通工具，各种电动自行车违法行为也随之出现；故对电动自行车违法行为进行识别，显得越来越重要。
目前，对电动自行车违法行为的识别，一般是通过对设置在被监控区域，比如路口、道路等的监控摄像头抓拍到的多张电动自行车的车辆图片进行图像识别，以确定电动自行车是否存在违法行为。然而，发明人意识到，若拍摄到的图片中的车辆较多，或者拍摄到的图片受天气等环境因素影响导致图片不清晰，很容易造成对电动自行车驾驶行为的误识别或者漏识别，进而导致电动自行车驾驶行为的识别准确率较低。
发明内容
根据本申请公开的各种实施例，提供一种电动自行车驾驶行为的识别方法、装置、计算机设备和存储介质。
一种电动自行车驾驶行为的识别方法包括：
接收双基识别设备发送的目标电动自行车行驶过被监控区域的行车视频以及车辆定位信息；所述双基识别设备用于监控所述被监控区域内的电动自行车的行车情况；
根据所述行车视频，从所述被监控区域对应的电子地图中确定出所述目标电动自行车的车辆位置，根据所述车辆位置确定所述目标电动自行车通过所述被监控区域的第一行车轨迹；
根据所述车辆定位信息确定所述目标电动自行车在所述电子地图中的定位位置，根据所述定位位置确定所述目标电动自行车通过所述被监控区域的第二行车轨迹；
根据所述第一行车轨迹和所述第二行车轨迹，确定所述目标电动自行车通过所述被监控区域的目标行车轨迹，根据所述目标行车轨迹确定所述目标电动自行车通过所述被监控区域的行车特征；及
根据所述行车特征确定对所述目标电动自行车的驾驶行为的识别结果。
一种电动自行车驾驶行为的识别装置包括：
信息接收模块，用于接收双基识别设备发送的目标电动自行车行驶过被监控区域的行车视频以及车辆定位信息；所述双基识别设备用于监控通过所述被监控区域内的电动自行车的行车情况；
第一行车轨迹确定模块，用于根据所述行车视频，从所述被监控区域对应的电子地图中确定出所述目标电动自行车的车辆位置，根据所述车辆位置确定所述目标电动自行车通过所述被监控区域的第一行车轨迹；
第二行车轨迹确定模块，用于根据所述车辆定位信息确定所述电动自行车在所述电子地图中的定位位置，根据所述定位位置确定所述电动自行车通过所述被监控区域的第二行车轨迹；
行车特征确定模块，用于根据所述第一行车轨迹和所述第二行车轨迹，确定所述目标电动自行车通过所述被监控区域的目标行车轨迹，根据所述目标行车轨迹确定所述电动自行车通过所述被监控区域的行车特征；及
识别结果确定模块，用于根据行车特征确定对所述目标电动自行车的驾驶行为的识别结果。
一种计算机设备，包括存储器和一个或多个处理器，所述存储器中储存有计算机可读指令，所述计算机可读指令被所述处理器执行时，使得所述一个或多个处理器执行以下步骤：
接收双基识别设备发送的目标电动自行车行驶过被监控区域的行车视频以及车辆定位信息；所述双基识别设备用于监控所述被监控区域内的电动自行车的行车情况；
根据所述行车视频，从所述被监控区域对应的电子地图中确定出所述目标电动自行车的车辆位置，根据所述车辆位置确定所述目标电动自行车通过所述被监控区域的第一行车轨迹；
根据所述车辆定位信息确定所述目标电动自行车在所述电子地图中的定位位置，根据所述定位位置确定所述目标电动自行车通过所述被监控区域的第二行车轨迹；
根据所述第一行车轨迹和所述第二行车轨迹，确定所述目标电动自行车通过所述被监控区域的目标行车轨迹，根据所述目标行车轨迹确定所述目标电动自行车通过所述被监控区域的行车特征；及
根据所述行车特征确定对所述目标电动自行车的驾驶行为的识别结果。
一个或多个存储有计算机可读指令的计算机可读存储介质，计算机可读指令被一个或多个处理器执行时，使得一个或多个处理器执行以下步骤：
接收双基识别设备发送的目标电动自行车行驶过被监控区域的行车视频以及车辆定位信息；所述双基识别设备用于监控所述被监控区域内的电动自行车的行车情况；
根据所述行车视频，从所述被监控区域对应的电子地图中确定出所述目标电动自行车的车辆位置，根据所述车辆位置确定所述目标电动自行车通过所述被监控区域的第一行车 轨迹；
根据所述车辆定位信息确定所述目标电动自行车在所述电子地图中的定位位置，根据所述定位位置确定所述目标电动自行车通过所述被监控区域的第二行车轨迹；
根据所述第一行车轨迹和所述第二行车轨迹，确定所述目标电动自行车通过所述被监控区域的目标行车轨迹，根据所述目标行车轨迹确定所述目标电动自行车通过所述被监控区域的行车特征；及
根据所述行车特征确定对所述目标电动自行车的驾驶行为的识别结果。
上述电动自行车驾驶行为的识别方法、装置、计算机设备和存储介质，综合考虑电动自行车的车辆定位信息以及车辆视频，使得识别出的电动自行车的驾驶行为更加准确，避免了仅仅基于拍摄到的图片对电动自行车驾驶行为进行识别，容易存在误识别或者漏识别，导致电动自行车驾驶行为的识别准确率较低的缺陷，进一步提高了电动自行车驾驶行为的识别准确率。
本申请的一个或多个实施例的细节在下面的附图和描述中提出。本申请的其它特征和优点将从说明书、附图以及权利要求书变得明显。
附图说明
图1为根据一个或多个实施例中电动自行车驾驶行为的识别方法的应用场景图；
图2为根据一个或多个实施例中电动自行车驾驶行为的识别方法的流程示意图；
图3为根据一个或多个实施例中十字交叉路口的示意图；
图4为根据一个或多个实施例中十字交叉路口的预设通行区域的示意图；
图5为另一个实施例中电动自行车驾驶行为的识别方法的流程示意图；
图6为根据一个或多个实施例中电动自行车驾驶行为的识别装置的框图；
图7为根据一个或多个实施例中计算机设备的框图。
具体实施方式
本申请提供的电动自行车驾驶行为的识别方法，可以应用于如图1所示的应用环境中。其中，双基识别设备110与服务器120通过网络进行通信。双基识别设备110设置在被监控区域，比如十字交叉道路的各个路口，用于实时监控通过被监控区域的电动自行车的行车情况，比如双基识别设备110采集目标电动自行车行驶过被监控区域的行车视频以及车辆定位信息，并将采集到的目标电动自行车行驶过被监控区域的的行车视频以及车辆定位信息发送至对应的服务器120。服务器120根据行车视频，从被监控区域对应的电子地图中确定出目标电动自行车的车辆位置，根据车辆位置确定目标电动自行车通过被监控区域的第一行车轨迹；根据车辆定位信息确定目标电动自行车在电子地图中的定位位置，根据定位位置确定目标电动自行车通过被监控区域的第二行车轨迹；根据第一行车轨迹和第二行车轨迹，确定目标电动自行车通过被监控区域的目标行车轨迹，根据目标行车轨迹 确定目标电动自行车通过被监控区域的行车特征；根据行车特征确定对目标电动自行车的驾驶行为的识别结果。其中，双基识别设备110是指安装有摄像装置(比如摄像头)的射频识别设备，比如双基RFID(Radio Frequency Identification，射频识别)设备；服务器120可以用独立的服务器或者是多个服务器组成的服务器集群来实现。
在其中一个实施例中，如图2所示，提供了一种电动自行车驾驶行为的识别方法，以该方法应用于图1中的服务器为例进行说明，包括以下步骤：
步骤S201，接收双基识别设备发送的目标电动自行车行驶过被监控区域的行车视频以及车辆定位信息；双基识别设备用于监控被监控区域内的电动自行车的行车情况。
在本步骤中，双基识别设备是指带有摄像头的RFID设备，比如双基RFID设备；双基识别设备能够监控被监控区域内的电动自行车的行车情况，并采集行驶过被监控区域的电动自行车的车辆信息、行车视频、车辆定位信息等；一般设置在路口处，比如十字交叉路口、T字形路口等。如图3所示，在十字交叉路口四个方向分别设置1台双基识别设备，并明确每台双基识别设备的天线朝向。当然，双基识别设备还可以设置在其他位置，比如道路旁边；具体设置位置根据被监控区域而定，本申请不做具体限定。车辆信息一般是指电动自行车的车牌信息、型号信息、车主信息等，行车视频是指通过双基识别设备中的摄像头拍摄得到的行驶过被监控区域的电动自行车的视频，车辆定位信息是指通过双基识别设备中的RFID设备采集得到的行驶过被监控区域的电动自行车的位置信息，可以是指电动自行车的经纬度坐标信息(如经度坐标信息和纬度坐标信息)，比如(31.2121751783,121.4411213954)；需要说明的是，本申请中的车辆定位信息是指多个车辆定位信息。
在本步骤中，电动自行车是指以蓄电池作为辅助能源的自行车，其安装有RFID标签，可以通过RFID标签写入设备向该RFID标签中写入电动自行车的车辆信息、年检信息、投保信息等；目标电动自行车是指需要识别驾驶行为的电动自行车；被监控区域是指双基识别设备的监控区域，比如路口、道路等。
具体实现中，设置在被监控区域附近的双基识别设备实时采集行驶过被监控区域的电动自行车的行车视频以及车辆定位信息，车辆视频和车辆定位信息均携带有车辆信息；从采集到的行驶过被监控区域的电动自行车的行车视频以及车辆定位信息中，筛选出属于同一车辆信息的行车视频以及车辆定位信息，作为行驶过被监控区域的目标电动自行车的行车视频以及车辆定位信息，并将行驶过被监控区域的目标电动自行车的行车视频以及车辆定位信息发送至对应的服务器；由服务器接收行驶过被监控区域的目标电动自行车的行车视频以及车辆定位信息，便于后续根据行驶过被监控区域的目标电动自行车的行车视频以及车辆定位信息，确定目标电动自行车的驾驶行为。
举例说明，参考图3，已安装有RFID标签的电动自行车行驶通过安装有双基RFID设备的路口，双基RFID设备发射超高频电磁波信号，电动自行车上安装的RFID标签接收超高频电磁波信号，启动芯片验证写入车辆信息，比如车牌信息，并将车辆信息发回至双基 RFID设备。同时，在这个过程中，双基RFID设备通过与电动自行车上安装的RFID标签进行多次通信，即可获取通过路口的电动自行车的多个车辆定位信息；或者，通过RFID定位技术，可以得到通过路口的电动自行车的多个车辆定位信息，每个车辆定位信息均携带有车辆信息，比如车牌信息。同时，双基RFID设备还可以启动摄像头拍摄通过上述路口的电动自行车的行车视频，并从行车视频中筛选出车辆关键帧静态图像，提取出车辆关键帧静态图像的图像特征，对图像特征进行分析，得到通过上述路口的电动自行车的车辆信息，比如车牌信息；其中，车辆关键帧静态图像是指能够清晰展示电动自行车的车牌号码的静态图像。接着，双基RFID设备从采集到的通过路口的电动自行车的行车视频以及车辆定位信息中，筛选出属于同一车牌信息的电动自行车的行车视频以及多个车辆定位信息，作为通过路口的目标电动自行车的行车视频以及车辆定位信息，并将通过路口的目标电动自行车的行车视频以及车辆定位信息发送至对应的服务器。
步骤S202，根据行车视频，从被监控区域对应的电子地图中确定出目标电动自行车的车辆位置，根据车辆位置确定目标电动自行车通过被监控区域的第一行车轨迹。
在本步骤中，电子地图是指将实际场景映射到计算机上的地图；被监控区域对应的电子地图是指将被监控区域的实际场景映射到计算机上的地图，如图3所示的十字交叉路口电子地图；通过被监控区域对应的电子地图有利于后续确定目标电动自行车的行车轨迹。第一行车轨迹是指基于目标电动自行车在被监控区域对应的电子地图中的车辆位置所确定的行车轨迹，能够反映电动自行车行驶过被监控区域的驾驶行为；比如，直行、向左转、向右转等，如图3所示。
具体地，服务器从行车视频中筛选出目标电动自行车的车辆关键帧静态图像，提取出车辆关键帧静态图像的图像特征，对图像特征进行分析，确定目标电动自行车在被监控区域所示图像中的位置信息；根据预设的被监控区域所示图像中的位置信息与被监控区域的电子地图中的位置的映射关系，确定目标电动自行车在被监控区域的电子地图中的车辆位置；将各个车辆位置作为位置节点进行连接，得到目标电动自行车通过被监控区域的第一行车轨迹。需要说明的是，由于双基识别设备的摄像头的拍摄角度固定，且覆盖整片被监控区域，那么行车视频中的视频帧对应的被监控区域所示图像的大小一致。
举例说明，服务器从行车视频中筛选出目标电动自行车的车辆关键帧静态图像，将车辆关键帧静态图像输入预先训练的卷积神经网络模型中，通过卷积神经网络模型提取车辆关键帧静态图像的图像特征，并对图像特征进行识别，得到目标电动自行车在被监控区域所示中的位置信息，比如对图像特征进行分析处理，得到目标电动自行车在被监控区域所示图像中的像素坐标，作为目标电动自行车在被监控区域所示图像中的位置信息；根据目标电动自行车在被监控区域所示图像中的位置信息，查询预设的被监控区域所示图像中的位置信息与被监控区域的电子地图中的位置的映射关系，确定目标电动自行车在被监控区域对应的电子地图中的车辆位置，比如确定目标电动自行车在被监控区域所示图像中的位置信息为位置A1，并根据预设的被监控区域所示图像中的位置信息与被监控区域的电子 地图中的位置的映射关系，确定与位置A1映射的位置为位置A2，则将位置A2作为目标电动自行车在被监控区域对应的电子地图中的车辆位置，以此类推，可以得到目标电动自行车在被监控区域对应的电子地图中的多个车辆位置；获取车辆关键帧静态图像的时间戳，作为对应确定出的车辆位置的时间信息；按照时间信息，将各个车辆位置作为位置节点进行连接，得到电动自行车通过被监控区域的第一行车轨迹。
步骤S203，根据车辆定位信息确定目标电动自行车在电子地图中的定位位置，根据定位位置确定目标电动自行车通过被监控区域的第二行车轨迹。
在本步骤中，第二行车轨迹是指基于目标电动自行车在被监控区域的电子地图中的定位位置所确定的行车轨迹，能够反映目标电动自行车通过被监控区域的驾驶行为；比如，直行、向左转、向右转等，如图3所示。
具体地，服务器预先去除重复的车辆定位信息，并根据预设的定位信息与路口的电子地图中的定位位置的映射关系，确定目标电动自行车在被监控区域的电子地图中的定位位置；将各个定位位置作为定位节点进行连接，得到目标电动自行车通过被监控区域的第二行车轨迹。
举例说明，服务器根据车辆定位信息查询预设的定位信息与被监控区域的电子地图中的定位位置的映射关系，确定目标电动自行车在被监控区域对应的电子地图中的定位位置；比如目标电动自行车的车辆定位信息为位置A3，则根据预设的定位信息与被监控区域的电子地图中的定位位置的映射关系，确定与位置A3映射的定位位置为位置A4，则将位置A4作为目标电动自行车在被监控区域对应的电子地图中的定位位置，以此类推，可以得到目标电动自行车在被监控区域对应的电子地图中的多个定位位置；获取车辆定位信息对应的时间信息，作为对应确定出的定位位置的时间信息；按照时间信息，将各个定位位置作为定位节点进行连接，得到目标电动自行车通过被监控区域的第二行车轨迹。
步骤S204，根据第一行车轨迹和第二行车轨迹，确定目标电动自行车通过被监控区域的目标行车轨迹，根据目标行车轨迹确定目标电动自行车通过被监控区域的行车特征。
在本步骤中，目标行车轨迹是指基于第一行车轨迹和第二行车轨迹综合确定的最终行车轨迹，能够反映目标电动自行车通过被监控区域的驾驶行为；比如，直行、向左转、向右转等，如图3所示。目标电动自行车通过被监控区域的行车特征用于衡量目标电动自行车通过被监控区域的行车信息，可以是指行驶方向、行驶区域或者通行状态等；其中，行驶方向是指直行通过路口、逆行通过路口、左转通过路口等；行驶区域是指主干道、禁限行区域、次干道、机动车道等；通行状态是指目标电动自行车通过行驶方向对应的路口的状态，比如闯红灯、绿灯通行等；在实际场景中，行车特征是指直行绿灯通行特征、左转绿灯通行特征、直行闯红灯特征、左转闯红灯特征、逆行通过特征、超载载物特征、未佩戴头盔特征、上桥隧特征、占用主干道特征、驶入禁限行区域特征等。
具体地，服务器根据第二行车轨迹对第一行车轨迹进行修正，得到修正后的第一行车轨迹，作为目标电动自行车通过被监控区域的目标行车轨迹；比如将第二行车轨迹的后半 部分轨迹，与第一行车轨迹的前半部分轨迹连接起来，得到修正后的第一行车轨迹，作为目标电动自行车通过被监控区域的目标行车轨迹；或者，服务器从第一行车轨迹中提取出目标电动自行车在被监控区域包括的各个预设通行区域的第一车辆位置，以及从第二行车轨迹中提取出目标电动自行车在各个预设通行区域的第二车辆位置；获取各个预设通行区域的第一车辆位置和第二车辆位置之间的位置平均值，作为各个预设通行区域的车辆位置；将各个预设通行区域的车辆位置当作位置节点进行连接，得到目标电动自行车通过被监控区域的目标行车轨迹。根据目标行车轨迹，确定电动自行车的行驶方向、行驶区域和通行状态，作为目标电动自行车通过被监控区域的行车特征。这样，综合考虑目标电动自行车的第一行车轨迹以及第二行车轨迹，使得后续识别出的目标电动自行车的驾驶行为更加准确，从而提高了电动自行车的驾驶行为的识别准确率。
举例说明，参考图3，服务器根据目标电动自行车的第一行车轨迹，确定电动自行车是直行通过路口，且当前交通灯状态为红灯；根据目标电动自行车的第二行车轨迹，确定电动自行车也是直行通过路口；则确认目标电动自行车通过路口的目标行车轨迹为直行闯红灯，且行车特征为直行、闯红灯；又例如，服务器根据目标电动自行车的第一行车轨迹，确定目标电动自行车是左转通过路口，且当前交通灯状态为红灯；根据目标电动自行车的第二行车轨迹，确定目标电动自行车也是左转通过路口，则确认目标电动自行车通过路口的目标行车轨迹为左转闯红灯，且行车特征为左转、闯红灯。
又例如，服务器根据目标电动自行车通过被监控区域的目标行车轨迹，确定目标电动自行车的驾驶方向为直行，驾驶区域为禁限行区域，则说明目标电动自行车通过被监控区域的行车特征为直行通过禁限行区域。
此外，服务器还可以根据行车视频中的全景大图，利用人体结构化、车辆结构化、人脸结构化算法识别目标电动自行车是否存在超载行为(比如超载载人、超载载物等)或者未佩戴头盔；同时根据双基RFID设备采集的车辆信息，确定目标电动自行车的车牌信息和车主信息；若目标电动自行车存在超载行为或者未佩戴头盔，则确认目标电动自行车通过被监控区域的行车特征为目标电动自行车超载或者目标电动自行车的车主未佩戴头盔。
步骤S205，根据行车特征确定对目标电动自行车的驾驶行为的识别结果。
具体地，服务器将行车特征与预设的违法行车特征(比如左转闯红灯、直行闯红灯等)等进行匹配；若匹配成功，则确认目标电动自行车通过被监控区域的驾驶行为为预设的违法驾驶行为；若匹配失败，则确认目标电动自行车通过被监控区域的驾驶行为不属于预设的违法驾驶行为，说明在通过被监控区域的过程中，目标电动自行车的驾驶行为为合法驾驶行为，从而得到对目标电动自行车的驾驶行为的识别结果。这样，根据目标电动自行车的目标行车轨迹，确定对目标电动自行车的驾驶行为的识别结果，避免了仅仅基于拍摄到的图片对电动自行车驾驶行为进行识别，容易存在误识别或者漏识别，导致电动自行车驾驶行为的识别准确率较低的缺陷，进一步提高了电动自行车驾驶行为的识别准确率；同时，整个过程中无需执行人员参与，从而实现了自动对通过被监控区域的电动自行车的驾驶行 为进行识别的目的，进而节约了大量人力成本，从而提高了对通过被监控区域的电动自行车的驾驶行为的识别效率。
上述电动自行车驾驶行为的识别方法中，实现了根据电动自行车的车辆定位信息以及行车视频，自动对电动自行车的驾驶行为进行识别的目的，结合基于行车视频确定的第一行车轨迹以及基于车辆定位信息确定的第二行车轨迹，确定电动自行车的目标行车轨迹，进而确定对电动自行车的驾驶行为的识别结果，避免了仅仅基于拍摄到的图片对电动自行车驾驶行为进行识别，容易存在误识别或者漏识别，导致电动自行车驾驶行为的识别准确率较低的缺陷，进一步提高了电动自行车驾驶行为的识别准确率。同时，综合考虑电动自行车的车辆定位信息以及车辆视频，使得识别出的电动自行车的驾驶行为更加准确，从而提高了电动自行车的驾驶行为的识别准确率。
在其中一个实施例中，车辆定位信息包括对应的时间信息，那么上述步骤S203，根据车辆定位信息确定目标电动自行车在电子地图中的定位位置，根据定位位置确定目标电动自行车通过被监控区域的第二行车轨迹，包括：获取各个车辆定位信息在电子地图中的位置标识符；根据位置标识符，确定目标电动自行车在电子地图中的定位位置；按照时间信息，将各个定位位置作为定位节点进行连接，得到目标电动自行车通过被监控区域的第二行车轨迹。
其中，位置标识符用于标识车辆定位信息在被监控区域的电子地图中的位置。
具体地，服务器获取各个车辆定位信息在被监控区域的电子地图中的位置标识符；根据位置标识符，确定各个车辆定位信息在被监控区域的电子地图中的位置，作为目标电动自行车在电子地图中的定位位置；按照车辆定位信息对应的时间信息，将对应确定出的定位位置当作定位节点依次连接起来，得到一条连接各个定位节点的线段，作为目标电动自行车通过被监控区域的第二行车轨迹。
在本实施例中，根据车辆定位信息，确定目标电动自行车通过被监控区域的第二行车轨迹，有利于后续结合电动自行车通过被监控区域的第一行车轨迹，自动对电动自行车的驾驶行为进行识别，避免了仅仅基于拍摄到的图片对电动自行车驾驶行为进行识别，容易存在误识别或者漏识别，导致电动自行车驾驶行为的识别准确率较低的缺陷，进一步提高了电动自行车驾驶行为的识别准确率。
在其中一个实施例中，上述步骤S204，根据第一行车轨迹和第二行车轨迹，确定目标电动自行车通过被监控区域的目标行车轨迹，包括：从第一行车轨迹中提取出目标电动自行车在被监控区域包括的预设通行区域的第一车辆位置；从第二行车轨迹中提取出目标电动自行车在预设通行区域的第二车辆位置；根据第一车辆位置和第二车辆位置，确定目标电动自行车在预设通行区域的车辆位置；将各个预设通行区域的车辆位置当作位置节点进行连接，得到目标电动自行车通过被监控区域的目标行车轨迹。
在本步骤中，被监控区域包括的预设通行区域是指由被监控区域划分而成的多个区域，具体是指被监控区域中的各个通行区域，比如将整个十字交叉路口划分成多个预设通 行区域，如图4所示的灰色小方格区域。
具体地，服务器识别第一行车轨迹经过的被监控区域中的预设通行区域，比如预设通行区域A、预设通行区域B、预设通行区域C等；从第一行车轨迹中提取出目标电动自行车在预设通行区域的第一车辆位置，比如从第一行车轨迹中提取出目标电动自行车在预设通行区域A、预设通行区域B、预设通行区域C的一个车辆位置，分别作为目标电动自行车在预设通行区域A、预设通行区域B、预设通行区域C的第一车辆位置；参照上述方式，从第二行车轨迹中提取出目标电动自行车在预设通行区域的第二车辆位置；获取目标电动自行车在预设通行区域的第一车辆位置和第二车辆位置之间的位置平均值，作为目标电动自行车在该预设通行区域的车辆位置；比如目标电动自行车在预设通行区域A的第一车辆位置为(a，b),第二车辆位置为(c，d)，则目标电动自行车在预设通行区域A的车辆位置为[(a+c)/2，(b+d)/2]；以此类推，可以得到目标电动自行车在多个预设通行区域的车辆位置；将各个预设通行区域的车辆位置当作位置节点进行连接，得到一条连接各个位置节点的线段，作为目标电动自行车通过被监控区域的目标行车轨迹。
在本实施例中，根据第一行车轨迹和第二行车轨迹，确定目标电动自行车通过路口的目标行车轨迹，有利于后续根据目标行车轨迹准确确定出目标电动自行车通过被监控区域的驾驶行为，从而提高了电动自行车驾驶行为的识别准确率。
在其中一个实施例中，上述步骤S204，根据目标行车轨迹确定电动自行车通过被监控区域的行车特征，包括：根据目标行车轨迹，确定目标电动自行车的行驶方向和行驶区域；获取与行驶方向对应的路口的当前交通灯状态，根据当前交通灯状态，确定目标电动自行车的通行状态；将目标电动自行车的行驶方向、行驶区域和通行状态，识别为目标电动自行车通过被监控区域的行车特征。
具体地，服务器对目标行车轨迹进行分析，得到电动自行车的行驶方向和行驶区域，从被监控区域关联的交通灯管理系统中获取与行驶方向对应的路口的当前交通灯状态，比如与直行路口对应的当前交通灯状态；根据当前交通灯状态，确定目标电动自行车的通行状态，比如若当前交通灯为红灯，说明目标电动自行车的通行状态为闯红灯；若当前交通灯为绿灯，说明目标电动自行车的通行状态为绿灯通行；将确定出的目标电动自行车的行驶方向、行驶区域和通行状态，识别为目标电动自行车通过被监控区域的行车特征。这样，有利于后续基于目标电动自行车通过被监控区域的行车特征，对目标电动自行车的驾驶行为进行识别，避免了仅仅基于拍摄到的图片对电动自行车驾驶行为进行识别，容易存在误识别或者漏识别，导致电动自行车驾驶行为的识别准确率较低的缺陷，进一步提高了电动自行车驾驶行为的识别准确率。
在其中一个实施例中，上述步骤S205，根据行车特征确定对目标电动自行车的驾驶行为的识别结果，包括：若行车特征与预设的违法行车特征匹配，则确认电动自行车的驾驶行为为违法驾驶行为；若行车特征与预设的违法行车特征不匹配，则确认电动自行车的驾驶行为为合法驾驶行为。
在本步骤中，行车特征用于表征目标电动自行车通过被监控区域的行车信息，比如直行绿灯通过特征、左转绿灯通过特征、直行闯红灯特征、左转闯红灯特征、超载载物特征、未佩戴头盔特征等；预设的违法行车特征是指从预设的违法驾驶行为信息中所提取出的行车特征，比如直行闯红灯特征、左转闯红灯特征、超载载物特征等。
具体地，服务器将行车特征与预设的违法行车特征进行匹配，若匹配成功，则确认目标电动自行车通过被监控区域的行车特征为预设的违法行车特征，说明目标电动自行车的驾驶行为为违法驾驶行为；若匹配失败，则确认目标电动自行车通过被监控区域的行车特征不是预设的违法行车特征，说明目标电动自行车的驾驶行为为合法驾驶行为。
举例说明，若行驶区域为主干道，说明目标电动自行车存在占用主干道的违法驾驶行为；若行驶方向为直行，通行状态为闯红灯等，说明目标电动自行车存在直行闯红灯的违法驾驶行为。这样，整个过程中无需执行人员参与，从而实现了自动对通过被监控区域的目标电动自行车的驾驶行为进行识别的目的，进而节约了大量人力成本，进一步提高了对通过路口的目标电动自行车的驾驶行为的识别效率。
在其中一个实施例中，行车视频携带有目标电动自行车的车辆信息；那么，在确认目标电动自行车的驾驶行为为违法驾驶行为之后，还包括：根据目标电动自行车的车辆信息，查询预设的车辆信息与车主信息的对应关系，得到目标电动自行车的车主信息；获取预设的提醒信息；将预设的提醒信息发送至车主信息绑定的终端。
在本步骤中，车主信息是指用于标识车主身份的信息，比如车主姓名、车主身份证号码等；车主信息绑定的终端可以是手机、平板电脑等。以车辆信息为车牌信息为例进行说明，服务器根据目标电动自行车的车牌信息，查询预设的车牌信息与车主信息的对应关系，得到目标电动自行车的车主信息；根据目标电动自行车的违法驾驶行为，从本地数据库中获取对应的预设提醒信息，比如“请不要闯红灯，注意安全驾驶”；将预设提醒信息通过短信的形式发送至车主信息对应的车主手机，以提醒车主安全驾驶。
在本实施例中，在确认目标电动自行车的驾驶行为为违法驾驶行为之后，通过将预设的提醒信息发送至车主信息绑定的终端，有利于提醒车主安全驾驶，避免出现安全事故。
在其中一个实施例中，在确认目标电动自行车的驾驶行为为违法驾驶行为之后，还包括：将目标电动自行车的违法驾驶行为上传至违章处理平台；违章处理平台用于根据目标电动自行车的违法驾驶行为，执行对应的违章处理操作。
具体地，在确认目标电动自行车的驾驶行为为违法驾驶行为之后，服务器还可以将目标电动自行车的违法驾驶行为上传至违章处理平台；违章处理平台自动根据目标电动自行车的违法驾驶行为，执行对应的违章处理操作，比如罚款、扣分等。当然，违章处理平台还可以存储记录目标电动自行车的违法驾驶行为，以进行备份和追溯。
在其中一个实施例中，本申请的电动自行车驾驶行为的识别方法还包括：根据对目标电动自行车的驾驶行为的识别结果，确定目标电动自行车的驾驶行为分数；根据预设的驾驶行为分数与驾驶行为安全等级的对应关系，确定目标电动自行车的驾驶行为安全等级。
在本步骤中，目标电动自行车的驾驶行为分数以及驾驶行为安全等级均用于衡量目标电动自行车对应的驾驶员的驾驶行为的安全程度，若驾驶行为分数较低，则对应的驾驶员的驾驶行为安全等级较低，说明驾驶员的驾驶行为存在危险性。例如，服务器对目标电动自行车的驾驶行为的识别结果，确定目标电动自行车的驾驶行为；查询预设的驾驶行为与驾驶行为分数的对应关系，得到目标电动自行车的驾驶行为分数；当目标电动自行车的驾驶行为分数大于或者等于第一阈值时，确定目标电动自行车的驾驶行为安全等级为第一安全等级；当目标电动自行车的驾驶行为分数大于第二阈值且小于第一阈值，确定目标电动自行车的驾驶行为安全等级为第二安全等级；当目标电动自行车的驾驶行为分数小于或等于第二阈值，确定目标电动自行车的驾驶行为安全等级为第三安全等级。
进一步地，在确定目标电动自行车的驾驶行为安全等级之后，服务器还可以将预设时间范围内(比如一个季度、一年等)的目标电动自行车的驾驶行为安全等级发送至车险服务器；车险服务器根据预设时间范围内的目标电动自行车的驾驶行为安全等级，确定目标电动自行车的驾驶行为综合等级；比如，获取预设时间范围内的目标电动自行车的驾驶行为安全等级的平均值，作为目标电动自行车的驾驶行为综合等级；获取与驾驶行为综合等级对应的车险报价信息，并将车险报价信息发送至车主信息绑定的终端，以供车主查看。例如，若驾驶行为综合等级较高，说明车主日常的驾驶行为都很好，则可以降低下一年的车险报价；若驾驶行为综合等级较低，说明车主日常的驾驶行为不是很好，则可以提高下一年的车险报价。
在其中一个实施例中，当目标电动自行车驶离路口时，通过RFID感知监控报警设备检测到目标电动自行车驶入不符合电动自行车充电的区域，诸如住宅小区大堂、楼道、架空层等，则触发报警，比如广播报警信息、开启报警灯等；并将报警信息发送至上述区域的工作人员，以提醒工作人员当前有电动自行车驶入不符合电动自行车充电的区域，请及时前往现场处理。
在其中一个实施例中，如图5所示，提供了另一种电动自行车驾驶行为的识别方法，以该方法应用于图1中的服务器为例进行说明，包括以下步骤：
步骤S501，接收双基识别设备发送的目标电动自行车行驶过被监控区域的行车视频以及车辆定位信息；双基识别设备用于监控被监控区域内的电动自行车的行车情况。
步骤S502，根据行车视频，从被监控区域对应的电子地图中确定出目标电动自行车的车辆位置，根据车辆位置确定目标电动自行车通过被监控区域的第一行车轨迹。
步骤S503，根据车辆定位信息确定目标电动自行车在电子地图中的定位位置，根据定位位置确定目标电动自行车通过被监控区域的第二行车轨迹。
步骤S504，根据第一行车轨迹和第二行车轨迹，确定目标电动自行车通过被监控区域的目标行车轨迹，根据目标行车轨迹确定目标电动自行车通过被监控区域的行车特征。
步骤S505，若行车特征与预设的违法行车特征匹配，则确认目标电动自行车的驾驶行为为违法驾驶行为。
步骤S506，根据目标电动自行车的车辆信息，查询预设的车辆信息与车主信息的对应关系，得到目标电动自行车的车主信息；获取预设的提醒信息；将预设的提醒信息发送至车主信息绑定的终端。
步骤S507，将目标电动自行车的违法驾驶行为上传至违章处理平台；违章处理平台用于根据目标电动自行车的违法驾驶行为，执行对应的违章处理操作。
步骤S508，根据对目标电动自行车的驾驶行为的识别结果，确定目标电动自行车的驾驶行为分数；根据预设的驾驶行为分数与驾驶行为安全等级的对应关系，确定目标电动自行车的驾驶行为安全等级。
上述电动自行车驾驶行为的识别方法，综合考虑电动自行车的车辆定位信息以及车辆视频，使得识别出的电动自行车的驾驶行为更加准确，避免了仅仅基于拍摄到的图片对电动自行车驾驶行为进行识别，容易存在误识别或者漏识别，导致电动自行车驾驶行为的识别准确率较低的缺陷，进一步提高了电动自行车驾驶行为的识别准确率。
应该理解的是，虽然图2、5的流程图中的各个步骤按照箭头的指示依次显示，但是这些步骤并不是必然按照箭头指示的顺序依次执行。除非本文中有明确的说明，这些步骤的执行并没有严格的顺序限制，这些步骤可以以其它的顺序执行。而且，图2、5中的至少一部分步骤可以包括多个子步骤或者多个阶段，这些子步骤或者阶段并不必然是在同一时刻执行完成，而是可以在不同的时刻执行，这些子步骤或者阶段的执行顺序也不必然是依次进行，而是可以与其它步骤或者其它步骤的子步骤或者阶段的至少一部分轮流或者交替地执行。
在其中一个实施例中，如图6所示，提供了一种电动自行车驾驶行为的识别装置，包括：信息接收模块610、第一行车轨迹确定模块620、第二行车轨迹确定模块630、行车特征确定模块640和识别结果确定模块650，其中：
信息接收模块610，用于接收双基识别设备发送的目标电动自行车行驶过被监控区域的行车视频以及车辆定位信息；双基识别设备用于监控通过被监控区域内的电动自行车的行车情况。
第一行车轨迹确定模块620，用于根据行车视频，从被监控区域对应的电子地图中确定出目标电动自行车的车辆位置，根据车辆位置确定目标电动自行车通过被监控区域的第一行车轨迹。
第二行车轨迹确定模块630，用于根据车辆定位信息确定目标电动自行车在电子地图中的定位位置，根据定位位置确定目标电动自行车通过被监控区域的第二行车轨迹。
行车特征确定模块640，用于根据第一行车轨迹和第二行车轨迹，确定目标电动自行车通过被监控区域的目标行车轨迹，根据目标行车轨迹确定目标电动自行车通过被监控区域的行车特征。
识别结果确定模块650，用于根据行车特征确定对目标电动自行车的驾驶行为的识别结果。
在其中一个实施例中，车辆定位信息包括对应的时间信息；第二行车轨迹确定模块630还用于获取各个车辆定位信息在电子地图中的位置标识符；根据位置标识符，确定目标电动自行车在电子地图中的定位位置；按照时间信息，将各个定位位置作为定位节点进行连接，得到目标电动自行车通过被监控区域的第二行车轨迹。
在其中一个实施例中，行车特征确定模块640还用于从第一行车轨迹中提取出目标电动自行车在被监控区域包括的预设通行区域的第一车辆位置；从第二行车轨迹中提取出目标电动自行车在预设通行区域的第二车辆位置；根据第一车辆位置和第二车辆位置，确定目标电动自行车在预设通行区域的车辆位置；将各个预设通行区域的车辆位置当作位置节点进行连接，得到目标电动自行车通过路口的目标行车轨迹；根据目标行车轨迹，确定目标电动自行车的行驶方向和行驶区域；获取与行驶方向对应的路口的当前交通灯状态，根据当前交通灯状态，确定目标电动自行车的通行状态；将目标电动自行车的行驶方向、行驶区域和通行状态，识别为目标电动自行车通过被监控区域的行车特征。
在其中一个实施例中，识别结果确定模块650还用于若行车特征与预设的违法行车特征匹配，则确认目标电动自行车的驾驶行为为违法驾驶行为；若行车特征与预设的违法行车特征不匹配，则确认目标电动自行车的驾驶行为为合法驾驶行为。
在其中一个实施例中，行车视频携带有目标电动自行车的车辆信息；电动自行车驾驶行为的识别装置还包括信息发送模块，用于根据目标电动自行车的车辆信息，查询预设的车辆信息与车主信息的对应关系，得到目标电动自行车的车主信息；获取预设的提醒信息；将预设的提醒信息发送至车主信息绑定的终端。
在其中一个实施例中，电动自行车驾驶行为的识别装置还包括上传模块，用于将目标电动自行车的违法驾驶行为上传至违章处理平台；违章处理平台用于根据目标电动自行车的违法驾驶行为，执行对应的违章处理操作。
在其中一个实施例中，电动自行车驾驶行为的识别装置还包括等级确定模块，用于根据对目标电动自行车的驾驶行为的识别结果，确定目标电动自行车的驾驶行为分数；根据预设的驾驶行为分数与驾驶行为安全等级的对应关系，确定目标电动自行车的驾驶行为安全等级。
上述各个实施例，综合考虑电动自行车的车辆定位信息以及车辆视频，使得识别出的电动自行车的驾驶行为更加准确，避免了仅仅基于拍摄到的图片对电动自行车驾驶行为进行识别，容易存在误识别或者漏识别，导致电动自行车驾驶行为的识别准确率较低的缺陷。
关于电动自行车驾驶行为的识别装置的具体限定可以参见上文中对于电动自行车驾驶行为的识别方法的限定，在此不再赘述。上述电动自行车驾驶行为的识别装置中的各个模块可全部或部分通过软件、硬件及其组合来实现。上述各模块可以硬件形式内嵌于或独立于计算机设备中的处理器中，也可以以软件形式存储于计算机设备中的存储器中，以便于处理器调用执行以上各个模块对应的操作。
在一个实施例中，提供了一种计算机设备，该计算机设备可以是服务器，其内部结构 图可以如图7所示。该计算机设备包括通过系统总线连接的处理器、存储器、网络接口和数据库。其中，该计算机设备的处理器用于提供计算和控制能力。该计算机设备的存储器包括非易失性或易失性存储介质、内存储器。该非易失性或易失性存储介质存储有操作系统、计算机可读指令和数据库。该内存储器为非易失性或易失性存储介质中的操作系统和计算机可读指令的运行提供环境。该计算机设备的数据库用于存储目标电动自行车的行车视频、车辆定位信息、行车轨迹、行车特征、驾驶行为的识别结果等数据。该计算机设备的网络接口用于与外部的终端通过网络连接通信。该计算机可读指令被处理器执行时以实现一种电动自行车驾驶行为的识别方法。
本领域技术人员可以理解，图7中示出的结构，仅仅是与本申请方案相关的部分结构的框图，并不构成对本申请方案所应用于其上的计算机设备的限定，具体的计算机设备可以包括比图中所示更多或更少的部件，或者组合某些部件，或者具有不同的部件布置。
一种计算机设备，包括存储器和一个或多个处理器，存储器中储存有计算机可读指令，计算机可读指令被处理器执行时实现本申请任意一个实施例中提供的电动自行车驾驶行为的识别方法的步骤。
一个或多个存储有计算机可读指令的计算机可读存储介质，所述计算机可读存储介质可以是非易失性，也可以是易失性，计算机可读指令被一个或多个处理器执行时，使得一个或多个处理器实现本申请任意一个实施例中提供的电动自行车驾驶行为的识别方法的步骤。
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分流程，是可以通过计算机可读指令来指令相关的硬件来完成，所述的计算机可读指令可存储于一非易失性或易失性计算机可读取存储介质中，该计算机可读指令在执行时，可包括如上述各方法的实施例的流程。其中，本申请所提供的各实施例中所使用的对存储器、存储、数据库或其它介质的任何引用，均可包括非易失性和易失性存储器中的至少一种。非易失性存储器可包括只读存储器(Read-Only Memory，ROM)、磁带、软盘、闪存或光存储器等。易失性存储器可包括随机存取存储器(Random Access Memory，RAM)或外部高速缓冲存储器。作为说明而非局限，RAM可以是多种形式，比如静态随机存取存储器(Static Random Access Memory，SRAM)或动态随机存取存储器(Dynamic Random Access Memory，DRAM)等。
以上实施例的各技术特征可以进行任意的组合，为使描述简洁，未对上述实施例中的各个技术特征所有可能的组合都进行描述，然而，只要这些技术特征的组合不存在矛盾，都应当认为是本说明书记载的范围。
以上所述实施例仅表达了本申请的几种实施方式，其描述较为具体和详细，但并不能因此而理解为对发明专利范围的限制。应当指出的是，对于本领域的普通技术人员来说，在不脱离本申请构思的前提下，还可以做出若干变形和改进，这些都属于本申请的保护范围。因此，本申请专利的保护范围应以所附权利要求为准。

Claims (22)

  1. 一种电动自行车驾驶行为的识别方法，包括：
    接收双基识别设备发送的目标电动自行车行驶过被监控区域的行车视频以及车辆定位信息；所述双基识别设备用于监控所述被监控区域内的电动自行车的行车情况；
    根据所述行车视频，从所述被监控区域对应的电子地图中确定出所述目标电动自行车的车辆位置，根据所述车辆位置确定所述目标电动自行车通过所述被监控区域的第一行车轨迹；
    根据所述车辆定位信息确定所述目标电动自行车在所述电子地图中的定位位置，根据所述定位位置确定所述目标电动自行车通过所述被监控区域的第二行车轨迹；
    根据所述第一行车轨迹和所述第二行车轨迹，确定所述目标电动自行车通过所述被监控区域的目标行车轨迹，根据所述目标行车轨迹确定所述目标电动自行车通过所述被监控区域的行车特征；及
    根据所述行车特征确定对所述目标电动自行车的驾驶行为的识别结果。
  2. 根据权利要求1所述的方法，其中，所述车辆定位信息包括对应的时间信息；
    所述根据所述车辆定位信息确定所述目标电动自行车在所述电子地图中的定位位置，根据所述定位位置确定所述目标电动自行车通过所述被监控区域的第二行车轨迹，包括：
    获取各个所述车辆定位信息在所述电子地图中的位置标识符；
    根据所述位置标识符，确定所述目标电动自行车在所述电子地图中的定位位置；及
    按照所述时间信息，将各个所述定位位置作为定位节点进行连接，得到所述目标电动自行车通过所述被监控区域的第二行车轨迹。
  3. 根据权利要求1所述的方法，其中，所述根据所述第一行车轨迹和所述第二行车轨迹，确定所述目标电动自行车通过所述被监控区域的目标行车轨迹，包括：
    从所述第一行车轨迹中提取出所述目标电动自行车在所述被监控区域包括的预设通行区域的第一车辆位置；
    从所述第二行车轨迹中提取出所述目标电动自行车在所述预设通行区域的第二车辆位置；
    根据所述第一车辆位置和所述第二车辆位置，确定所述目标电动自行车在所述预设通行区域的车辆位置；及
    将各个所述预设通行区域的车辆位置当作位置节点进行连接，得到所述目标电动自行车通过所述被监控区域的目标行车轨迹；
    所述根据所述目标行车轨迹确定所述目标电动自行车通过所述被监控区域的行车特征，包括：
    根据所述目标行车轨迹，确定所述目标电动自行车的行驶方向和行驶区域；
    获取与所述行驶方向对应的路口的当前交通灯状态，根据所述当前交通灯状态，确定所述目标电动自行车的通行状态；及
    将所述目标电动自行车的所述行驶方向、所述行驶区域和所述通行状态，识别为所述目标电动自行车通过所述被监控区域的行车特征。
  4. 根据权利要求1所述的方法，其中，所述根据所述行车特征确定对所述目标电动自行车的驾驶行为的识别结果，包括：
    若所述行车特征与预设的违法行车特征匹配，则确认所述目标电动自行车的驾驶行为为违法驾驶行为；及
    若所述行车特征与所述预设的违法行车特征不匹配，则确认所述目标电动自行车的驾驶行为为合法驾驶行为。
  5. 根据权利要求4所述的方法，其中，所述行车视频携带有所述目标电动自行车的车辆信息；
    在确认所述目标电动自行车的驾驶行为为违法驾驶行为之后，所述方法还包括：
    根据所述目标电动自行车的车辆信息，查询预设的车辆信息与车主信息的对应关系，得到所述目标电动自行车的车主信息；
    获取预设的提醒信息；及
    将所述预设的提醒信息发送至所述车主信息绑定的终端。
  6. 根据权利要求4所述的方法，其中，在确认所述目标电动自行车的驾驶行为为违法驾驶行为之后，所述方法还包括：
    将所述目标电动自行车的所述违法驾驶行为上传至违章处理平台；所述违章处理平台用于根据所述目标电动自行车的所述违法驾驶行为，执行对应的违章处理操作。
  7. 根据权利要求1至6任一项所述的方法，其中，还包括：
    根据对所述目标电动自行车的驾驶行为的识别结果，确定所述目标电动自行车的驾驶行为分数；及
    根据预设的驾驶行为分数与驾驶行为安全等级的对应关系，确定所述目标电动自行车的驾驶行为安全等级。
  8. 一种电动自行车驾驶行为的识别装置，包括：
    信息接收模块，用于接收双基识别设备发送的目标电动自行车行驶过被监控区域的行车视频以及车辆定位信息；所述双基识别设备用于监控通过所述被监控区域内的电动自行车的行车情况；
    第一行车轨迹确定模块，用于根据所述行车视频，从所述被监控区域对应的电子地图中确定出所述目标电动自行车的车辆位置，根据所述车辆位置确定所述目标电动自行车通过所述被监控区域的第一行车轨迹；
    第二行车轨迹确定模块，用于根据所述车辆定位信息确定所述电动自行车在所述电子地图中的定位位置，根据所述定位位置确定所述电动自行车通过所述被监控区域的第二行车轨迹；
    行车特征确定模块，用于根据所述第一行车轨迹和所述第二行车轨迹，确定所述目标 电动自行车通过所述被监控区域的目标行车轨迹，根据所述目标行车轨迹确定所述电动自行车通过所述被监控区域的行车特征；及
    识别结果确定模块，用于根据所述行车特征确定对所述电动自行车的驾驶行为的识别结果。
  9. 一种计算机设备，包括存储器及一个或多个处理器，所述存储器中储存有计算机可读指令，所述计算机可读指令被所述一个或多个处理器执行时，使得所述一个或多个处理器执行以下步骤：
    接收双基识别设备发送的目标电动自行车行驶过被监控区域的行车视频以及车辆定位信息；所述双基识别设备用于监控所述被监控区域内的电动自行车的行车情况；
    根据所述行车视频，从所述被监控区域对应的电子地图中确定出所述目标电动自行车的车辆位置，根据所述车辆位置确定所述目标电动自行车通过所述被监控区域的第一行车轨迹；
    根据所述车辆定位信息确定所述目标电动自行车在所述电子地图中的定位位置，根据所述定位位置确定所述目标电动自行车通过所述被监控区域的第二行车轨迹；
    根据所述第一行车轨迹和所述第二行车轨迹，确定所述目标电动自行车通过所述被监控区域的目标行车轨迹，根据所述目标行车轨迹确定所述目标电动自行车通过所述被监控区域的行车特征；及
    根据所述行车特征确定对所述目标电动自行车的驾驶行为的识别结果。
  10. 根据权利要求9所述的计算机设备，其中，所述车辆定位信息包括对应的时间信息，所述处理器执行所述计算机可读指令时还执行以下步骤：
    获取各个所述车辆定位信息在所述电子地图中的位置标识符；
    根据所述位置标识符，确定所述目标电动自行车在所述电子地图中的定位位置；及
    按照所述时间信息，将各个所述定位位置作为定位节点进行连接，得到所述目标电动自行车通过所述被监控区域的第二行车轨迹。
  11. 根据权利要求9所述的计算机设备，其中，所述处理器执行所述计算机可读指令时还执行以下步骤：
    从所述第一行车轨迹中提取出所述目标电动自行车在所述被监控区域包括的预设通行区域的第一车辆位置；
    从所述第二行车轨迹中提取出所述目标电动自行车在所述预设通行区域的第二车辆位置；
    根据所述第一车辆位置和所述第二车辆位置，确定所述目标电动自行车在所述预设通行区域的车辆位置；
    将各个所述预设通行区域的车辆位置当作位置节点进行连接，得到所述目标电动自行车通过所述被监控区域的目标行车轨迹；
    根据所述目标行车轨迹，确定所述目标电动自行车的行驶方向和行驶区域；
    获取与所述行驶方向对应的路口的当前交通灯状态，根据所述当前交通灯状态，确定所述目标电动自行车的通行状态；及
    将所述目标电动自行车的所述行驶方向、所述行驶区域和所述通行状态，识别为所述目标电动自行车通过所述被监控区域的行车特征。
  12. 根据权利要求9所述的计算机设备，其中，所述处理器执行所述计算机可读指令时还执行以下步骤：
    若所述行车特征与预设的违法行车特征匹配，则确认所述目标电动自行车的驾驶行为为违法驾驶行为；及
    若所述行车特征与所述预设的违法行车特征不匹配，则确认所述目标电动自行车的驾驶行为为合法驾驶行为。
  13. 根据权利要求12所述的计算机设备，其中，所述行车视频携带有所述目标电动自行车的车辆信息，所述处理器执行所述计算机可读指令时还执行以下步骤：
    根据所述目标电动自行车的车辆信息，查询预设的车辆信息与车主信息的对应关系，得到所述目标电动自行车的车主信息；
    获取预设的提醒信息；及
    将所述预设的提醒信息发送至所述车主信息绑定的终端。
  14. 根据权利要求12所述的计算机设备，其中，所述处理器执行所述计算机可读指令时还执行以下步骤：
    将所述目标电动自行车的所述违法驾驶行为上传至违章处理平台；所述违章处理平台用于根据所述目标电动自行车的所述违法驾驶行为，执行对应的违章处理操作。
  15. 根据权利要求9至14任一项所述的计算机设备，其中，所述处理器执行所述计算机可读指令时还执行以下步骤：
    根据对所述目标电动自行车的驾驶行为的识别结果，确定所述目标电动自行车的驾驶行为分数；及
    根据预设的驾驶行为分数与驾驶行为安全等级的对应关系，确定所述目标电动自行车的驾驶行为安全等级。
  16. 一个或多个存储有计算机可读指令的计算机可读存储介质，所述计算机可读指令被一个或多个处理器执行时，使得所述一个或多个处理器执行以下步骤：
    接收双基识别设备发送的目标电动自行车行驶过被监控区域的行车视频以及车辆定位信息；所述双基识别设备用于监控所述被监控区域内的电动自行车的行车情况；
    根据所述行车视频，从所述被监控区域对应的电子地图中确定出所述目标电动自行车的车辆位置，根据所述车辆位置确定所述目标电动自行车通过所述被监控区域的第一行车轨迹；
    根据所述车辆定位信息确定所述目标电动自行车在所述电子地图中的定位位置，根据所述定位位置确定所述目标电动自行车通过所述被监控区域的第二行车轨迹；
    根据所述第一行车轨迹和所述第二行车轨迹，确定所述目标电动自行车通过所述被监控区域的目标行车轨迹，根据所述目标行车轨迹确定所述目标电动自行车通过所述被监控区域的行车特征；及
    根据所述行车特征确定对所述目标电动自行车的驾驶行为的识别结果。
  17. 根据权利要求16所述的存储介质，其中，所述车辆定位信息包括对应的时间信息，所述计算机可读指令被所述处理器执行时还执行以下步骤：
    获取各个所述车辆定位信息在所述电子地图中的位置标识符；
    根据所述位置标识符，确定所述目标电动自行车在所述电子地图中的定位位置；及
    按照所述时间信息，将各个所述定位位置作为定位节点进行连接，得到所述目标电动自行车通过所述被监控区域的第二行车轨迹。
  18. 根据权利要求16所述的存储介质，其中，所述计算机可读指令被所述处理器执行时还执行以下步骤：
    从所述第一行车轨迹中提取出所述目标电动自行车在所述被监控区域包括的预设通行区域的第一车辆位置；
    从所述第二行车轨迹中提取出所述目标电动自行车在所述预设通行区域的第二车辆位置；
    根据所述第一车辆位置和所述第二车辆位置，确定所述目标电动自行车在所述预设通行区域的车辆位置；
    将各个所述预设通行区域的车辆位置当作位置节点进行连接，得到所述目标电动自行车通过所述被监控区域的目标行车轨迹；
    根据所述目标行车轨迹，确定所述目标电动自行车的行驶方向和行驶区域；
    获取与所述行驶方向对应的路口的当前交通灯状态，根据所述当前交通灯状态，确定所述目标电动自行车的通行状态；及
    将所述目标电动自行车的所述行驶方向、所述行驶区域和所述通行状态，识别为所述目标电动自行车通过所述被监控区域的行车特征。
  19. 根据权利要求16所述的存储介质，其中，所述计算机可读指令被所述处理器执行时还执行以下步骤：
    若所述行车特征与预设的违法行车特征匹配，则确认所述目标电动自行车的驾驶行为为违法驾驶行为；及
    若所述行车特征与所述预设的违法行车特征不匹配，则确认所述目标电动自行车的驾驶行为为合法驾驶行为。
  20. 根据权利要求19所述的存储介质，其中，所述行车视频携带有所述目标电动自行车的车辆信息，所述计算机可读指令被所述处理器执行时还执行以下步骤：
    根据所述目标电动自行车的车辆信息，查询预设的车辆信息与车主信息的对应关系，得到所述目标电动自行车的车主信息；
    获取预设的提醒信息；及
    将所述预设的提醒信息发送至所述车主信息绑定的终端。
  21. 根据权利要求19所述的存储介质，其中，所述计算机可读指令被所述处理器执行时还执行以下步骤：
    将所述目标电动自行车的所述违法驾驶行为上传至违章处理平台；所述违章处理平台用于根据所述目标电动自行车的所述违法驾驶行为，执行对应的违章处理操作。
  22. 根据权利要求16至21任一项所述的存储介质，其中，所述计算机可读指令被所述处理器执行时还执行以下步骤：
    根据对所述目标电动自行车的驾驶行为的识别结果，确定所述目标电动自行车的驾驶行为分数；及
    根据预设的驾驶行为分数与驾驶行为安全等级的对应关系，确定所述目标电动自行车的驾驶行为安全等级。
PCT/CN2020/098383 2020-03-16 2020-06-28 电动自行车驾驶行为的识别方法、装置和计算机设备 WO2021184570A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010181183.7A CN111444798B (zh) 2020-03-16 2020-03-16 电动自行车驾驶行为的识别方法、装置和计算机设备
CN202010181183.7 2020-03-16

Publications (1)

Publication Number Publication Date
WO2021184570A1 true WO2021184570A1 (zh) 2021-09-23

Family

ID=71648885

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/098383 WO2021184570A1 (zh) 2020-03-16 2020-06-28 电动自行车驾驶行为的识别方法、装置和计算机设备

Country Status (2)

Country Link
CN (1) CN111444798B (zh)
WO (1) WO2021184570A1 (zh)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113888883A (zh) * 2021-09-29 2022-01-04 北京无线电计量测试研究所 一种自行车交通智能监测系统和方法
CN114241774A (zh) * 2021-12-28 2022-03-25 高新兴智联科技有限公司 一种记录车辆违法证据方法和装置
CN115410381A (zh) * 2022-06-29 2022-11-29 合肥天互光电科技有限公司 一种基于人脸识别的非机动车交通管制用监控提醒装置

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112016653B (zh) * 2020-09-04 2023-04-07 深圳华强技术有限公司 一种检测驾驶员是否佩戴头盔的系统和方法
CN113920722B (zh) * 2021-09-23 2023-04-14 摩拜(北京)信息技术有限公司 路口通行状态获取方法、装置、电子设备及存储介质
CN115909286B (zh) * 2023-02-17 2023-07-04 浙江口碑网络技术有限公司 数据处理方法及装置
CN116994178A (zh) * 2023-07-31 2023-11-03 赛力斯汽车有限公司 视频处理方法、视频识别方法、视频处理装置以及车辆

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050219375A1 (en) * 2004-03-31 2005-10-06 Makoto Hasegawa Method of retrieving image data of a moving object, apparatus for photographing and detecting a moving object, and apparatus for retrieving image data of a moving object
CN106448181A (zh) * 2016-10-31 2017-02-22 广东中星电子有限公司 非机动车管理方法、装置及系统
US9595192B1 (en) * 2016-04-14 2017-03-14 Ahmad Abdulaziz Alrashid Traffic violation detection system
CN107122883A (zh) * 2017-03-22 2017-09-01 北京零科技有限公司 一种基于射频识别与智能视频技术的电动车管控系统
CN109743686A (zh) * 2018-12-27 2019-05-10 福建小电科技有限公司 一种基于大数据的电动自行车管理系统
CN109949582A (zh) * 2019-04-25 2019-06-28 深圳市华士精成科技有限公司 一种基于无线定位的车辆违法行驶监控系统

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105844914B (zh) * 2016-04-25 2019-01-25 深圳市双赢伟业科技股份有限公司 路况监控方法和装置
CN110178167B (zh) * 2018-06-27 2022-06-21 潍坊学院 基于摄像机协同接力的路口违章视频识别方法
CN109389827A (zh) * 2018-08-17 2019-02-26 深圳壹账通智能科技有限公司 基于行车记录仪的举证方法、装置、设备和存储介质
CN108847032A (zh) * 2018-08-21 2018-11-20 北京深瞐科技有限公司 一种交通违法行为识别方法及装置
CN110097759A (zh) * 2019-04-28 2019-08-06 南京师范大学 一种基于视频地理围栏的机动车违章行为检测方法

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050219375A1 (en) * 2004-03-31 2005-10-06 Makoto Hasegawa Method of retrieving image data of a moving object, apparatus for photographing and detecting a moving object, and apparatus for retrieving image data of a moving object
US9595192B1 (en) * 2016-04-14 2017-03-14 Ahmad Abdulaziz Alrashid Traffic violation detection system
CN106448181A (zh) * 2016-10-31 2017-02-22 广东中星电子有限公司 非机动车管理方法、装置及系统
CN107122883A (zh) * 2017-03-22 2017-09-01 北京零科技有限公司 一种基于射频识别与智能视频技术的电动车管控系统
CN109743686A (zh) * 2018-12-27 2019-05-10 福建小电科技有限公司 一种基于大数据的电动自行车管理系统
CN109949582A (zh) * 2019-04-25 2019-06-28 深圳市华士精成科技有限公司 一种基于无线定位的车辆违法行驶监控系统

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113888883A (zh) * 2021-09-29 2022-01-04 北京无线电计量测试研究所 一种自行车交通智能监测系统和方法
CN114241774A (zh) * 2021-12-28 2022-03-25 高新兴智联科技有限公司 一种记录车辆违法证据方法和装置
CN114241774B (zh) * 2021-12-28 2023-09-19 高新兴智联科技股份有限公司 一种记录车辆违法证据方法和装置
CN115410381A (zh) * 2022-06-29 2022-11-29 合肥天互光电科技有限公司 一种基于人脸识别的非机动车交通管制用监控提醒装置

Also Published As

Publication number Publication date
CN111444798A (zh) 2020-07-24
CN111444798B (zh) 2024-05-28

Similar Documents

Publication Publication Date Title
WO2021184570A1 (zh) 电动自行车驾驶行为的识别方法、装置和计算机设备
CN108062349B (zh) 基于视频结构化数据及深度学习的视频监控方法和系统
US11481991B2 (en) System and method for detecting and transmitting incidents of interest of a roadway to a remote server
CN110738150B (zh) 相机联动抓拍方法、装置以及计算机存储介质
WO2021223418A1 (zh) 基于视觉差的停车检测方法及装置
CN109714576B (zh) 基于视频监控的车辆识别方法、装置、系统及服务器
CN110619277A (zh) 一种多社区智慧布控方法以及系统
WO2021227586A1 (zh) 交通事故分析方法、装置及设备
CN109816971B (zh) 基于多源数据融合的危险品运输车辆预防跟踪系统及方法
CN111079751B (zh) 识别车牌真伪的方法、装置、计算机设备及存储介质
CN105528626A (zh) 一种rfid读写器和摄像机集成一体机及其应用
CN112084892B (zh) 一种道路异常事件检测管理装置及其方法
CN113903008A (zh) 基于深度学习和轨迹跟踪的匝道出口车辆违法识别方法
CN111860219B (zh) 一种高速占道判断方法、装置及电子设备
CN108847031A (zh) 交通行为监控方法、装置、计算机设备和存储介质
US20100027841A1 (en) Method and system for detecting a signal structure from a moving video platform
CN111078973A (zh) 基于大数据的套牌车识别方法、设备及存储介质
CN113723273A (zh) 一种车辆轨迹信息确定方法、装置及计算机设备
CN114141022B (zh) 应急车道占用行为检测方法、装置、电子设备及存储介质
CN116912517B (zh) 相机视野边界的检测方法及装置
CN108510757A (zh) 一种基于图像识别的十字路口智能交通灯管理系统及方法
US20160358462A1 (en) Method and system for vehicle data integration
CN115861916A (zh) 异常停车行为检测方法、装置、计算机设备和存储介质
CN114693722A (zh) 一种车辆行驶行为检测方法、检测装置及检测设备
TW202341006A (zh) 物件追蹤整合方法及整合裝置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20925633

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 11/01/2023)

122 Ep: pct application non-entry in european phase

Ref document number: 20925633

Country of ref document: EP

Kind code of ref document: A1