Method and System for Warning of Traffic Accidents in Column Blind Zones

Abstract

The present invention provides a method and system for warning of traffic accidents in a blind area of a column. When the first road where the blind area of the column is located is in a red light state, a camera is used to collect a video of the blind area of the column, and the video is input into a target detection model to detect whether there is a target to be measured in the blind area of the column; a coordinate system is constructed for the blind area of the column in the video. When a target to be measured is detected in the blind area of the column, the target to be measured is located through the video and a sensor respectively, and the speed and acceleration of the target to be measured are calculated; according to the speed, acceleration of the target to be measured and the distance from the second road, the danger level is comprehensively evaluated, and vehicles on the second road are warned differently under different danger levels. The present invention accurately predicts the danger level of the target by positioning the target and calculating the speed and acceleration, effectively avoiding traffic accidents caused by pedestrians in the blind area and ensuring the safety of traffic participants.

Description

Technical Field

[0001] The present invention relates to the field of intelligent transportation, and specifically, to a method and system for warning of accidents in the blind area of columns. Background Art

[0002] Elevated roads are mainly built to improve driving speed or solve the safety problems of the intersection of roads with railways or pedestrian traffic lines. They are connected to ground roads through ramps and avoid intersections with ground roads in a grade-separated manner. In urban areas, in order to meet the need for rapid access to the urban central area, there are also designs where roads tens of kilometers long are all on elevated roads.

[0003] Elevated roads not only bring great convenience to urban commuting, but also pose certain traffic safety hazards. For example, Figure 1 As shown, for vehicles traveling from an intersection towards the main road, the area directly behind Column 1 and the area behind the right side of Column 2 are blind spots in the field of vision, and this area is exactly the location of the non-motor vehicle lane and the zebra crossing. If a non-motor vehicle or a pedestrian rushes out of the blind spot when the vehicle passes through the intersection normally under a green light, it is very likely to cause a traffic accident.

[0004] After retrieval, it is found that the existing research mainly focuses on warning of traffic accidents in the right-turn blind area of vehicles themselves, such as patent document CN117912257A. For the research on warning of traffic accidents in the blind area behind elevated columns, it mainly judges whether there are pedestrians and non-motor vehicles by detecting whether there are moving objects in the blind area through motion sensors. Although this judgment method can detect pedestrians and non-motor vehicles, it can only indicate that there are people in the blind area and cannot predict whether pedestrians and non-motor vehicles will rush out of the blind area, so the safety is limited. Summary of the Invention

[0005] Aiming at the deficiencies in the prior art, the purpose of the present invention is to provide a method and system for warning of traffic accidents in the blind area of columns.

[0006] According to a method for warning of traffic accidents in the blind area of columns provided by the present invention, it includes:

[0007] Target recognition step: When the first road where the blind area of the column is located is red, collect the video of the blind area of the column through a camera, input the video into a target detection model, and detect whether there is a target to be detected in the blind area of the column;

[0008] Moving state calculation step: Establish a coordinate system for the blind area of the column in the video. When a target to be detected is detected in the blind area of the column, locate the target to be detected through the video and sensors respectively, and calculate the speed and acceleration of the target to be detected;

[0009] Decision-making step: Based on the speed, acceleration of the target to be measured and the distance from the second road, comprehensively evaluate to obtain the degree of danger, and issue different warnings to the vehicles on the second road under different degrees of danger.

[0010] Furthermore, the methods for positioning the target to be measured include:

[0011] Determine the first position information (x1, y1, z1) of the target to be measured according to its position in the video, and the standard deviation of its position is (σ x1 , σ y1 , σ z1 ). Determine the second position information (x2, y2, z2) of the target to be measured through a millimeter-wave radar, and the standard deviation of its position is (σ x2 , σ y2 , σ z2 ). Determine the third position information (x3, y3, z3) of the target to be measured through a laser sensor, and the standard deviation of the position is (σ x3 , σ y3 , σ z3 ). Through the Bayesian fusion algorithm, obtain the fused position information (x, y, z) and the standard deviation (σ x , σ y , σ z ).

[0012] Furthermore, for the x coordinate:

[0013] , ,

[0014] K = + +

[0015]

[0016]

[0017] The same applies to the y coordinate and the z coordinate.

[0018] Furthermore, the methods for calculating the speed of the target to be measured include:

[0019] Obtain the fused position information (x t1 , y t1 , z t1 ) of the target to be measured at time t1, and the fused position information (x t2 , y t2 , z t2 ) at time t2, Δt = t2 - t1;

[0020] The speed of the target to be measured :

[0021] 。

[0022] Furthermore, the calculation method of the acceleration of the target to be measured includes:

[0023] Obtain the fused position information (x t3 , y t3 , z t3 ) of the target to be measured at time t3

[0024] The acceleration of the target to be measured in the x direction :

[0025]

[0026] The accelerations of the target to be measured in the y and z directions , Similarly.

[0027] Furthermore, the method for comprehensively evaluating the degree of danger includes:

[0028]

[0029] d is the perpendicular distance from the fused position information to the edge of the column blind area close to the second road.

[0030] Furthermore, when there is no target to be measured in the column blind area, D = 0, which is the first degree of danger, and the vehicles on the second road are warned by a green signal light. When D belongs to (0, D0), it is the second degree of danger, and the vehicles on the second road are warned by switching between red and yellow signal lights. When D is greater than or equal to D0, it is the third degree of danger, and the vehicles on the second road are warned by a red signal light and a projection lamp.

[0031] According to an accident warning system for column blind areas provided by the present invention, it includes:

[0032] Camera: When the first road where the column blind area is located has a red light, collect the video of the column blind area;

[0033] Target detection model: Obtain the video of the column blind area collected by the camera and detect whether there is a target to be measured in the column blind area;

[0034] Processing module: Construct a coordinate system for the blind area of the columns in the video. When a target to be measured is detected in the blind area of the columns, locate the target to be measured through the video and sensors respectively, calculate the speed and acceleration of the target to be measured, and comprehensively evaluate the degree of danger based on the speed, acceleration of the target to be measured and the distance from the second road. Under different degrees of danger, give different warnings to the vehicles on the second road.

[0035] Further, the sensor includes: a millimeter-wave radar and a laser sensor.

[0036] Further, it also includes a signal lamp and a projection lamp, and give warnings through the signal lamp and the projection lamp.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] The present invention accurately predicts the degree of danger of the target by positioning the target and calculating the speed and acceleration, effectively avoids traffic accidents caused by pedestrians in the blind area, makes up for the congenital deficiency of drivers and vehicles themselves in observing the blind area, and ensures the safety of traffic participants.

[0039] The present invention gives full play to the advantages of different sensors through multi-sensor fusion, improves the detection accuracy and reliability of pedestrians in the blind area behind the elevated columns, and can effectively detect pedestrians whether it is day, night or under bad weather conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] By reading the following detailed description of the non-limiting embodiments with reference to the accompanying drawings, other features, objects and advantages of the present invention will become more apparent:

[0041] Figure 1 It is a schematic diagram of the road under the viaduct;

[0042] Figure 2 It is the working flow chart of the present invention;

[0043] Figure 3 It is the test effect diagram of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0044] The present invention will be described in detail below with reference to specific embodiments. The following embodiments will help those skilled in the art to further understand the present invention, but do not limit the present invention in any form. It should be noted that those of ordinary skill in the art can make several changes and improvements without departing from the concept of the present invention. These all belong to the protection scope of the present invention.

[0045] As Figure 2 shown, according to a method for warning traffic accidents in the blind area of columns provided by the present invention, it includes:

[0046] Target recognition step: When the first road ( Figure 1 the non-motor vehicle lane therein) where the blind area of the column is located is red, the video of the blind area of the column is collected by the camera, and the video is input into the target detection model to detect whether there is a target to be measured in the blind area of the column. Since Figure 1 there are two columns, two blind areas will be generated for the second road (the main road part of the motor vehicle lane). Among them, the first column generates the first blind area, and the second column generates the second blind area. Therefore, separate detections are required.

[0047] The camera is used to collect image or video information of the blind area behind the elevated column, providing basic data for subsequent image analysis. A camera with high resolution and high frame rate can be selected to ensure clear images can be obtained under different lighting and weather conditions. The millimeter-wave radar can detect the distance, speed, and position information of pedestrians by transmitting and receiving millimeter-wave signals, and has advantages especially in bad weather or low light conditions, and can penetrate some obstacles for detection. The laser sensor can accurately measure the distance between the pedestrian and the column, and accurately judge the position information of the pedestrian by emitting a laser beam and receiving the reflected signal. In the present invention, it is preferably to turn on each sensor and start the subsequent calculation steps after detecting the target, so as to avoid waste of energy and computing resources. The installation position of each sensor can be on the Figure 1 back of the first column in

[0048] The target detection model uses a deep learning algorithm to process the image or video data collected by the camera, and recognizes pedestrians and non-motor vehicles therein. Here, deep learning architectures such as convolutional neural networks (CNNs) can be used. Through a pre-trained model, the features of pedestrians and non-motor vehicles in the image can be automatically extracted.

[0049] Motion state calculation step: A coordinate system is constructed for the blind area of the column in the video. When a target to be measured is detected in the blind area of the column, the target to be measured is located through the video and the sensor respectively, and the speed and acceleration of the target to be measured are calculated.

[0050] In this step, first, operations such as format conversion, denoising, and normalization can be performed on the data collected by the sensor to make the data meet the requirements of subsequent processing. Then, the data from different sensors are fused, and combined with the image recognition information and the distance and speed information of the radar and laser sensor, to comprehensively judge the accurate position and motion state of the pedestrian.

[0051] The methods for positioning the target to be measured include:

[0052] According to the position of the target to be measured in the video, the first position information (x1, y1, z1) of the target to be measured is determined, and the standard deviation of its position is (σ x1 σy1 , σ z1 ), the second position information (x2, y2, z2) of the target to be measured is determined by a millimeter-wave radar, and the standard deviation of its position is (σ x2 , σ y2 , σ z2 ). The third position information (x3, y3, z3) of the target to be measured is determined by a laser sensor, and the standard deviation of the position is (σ x3 , σ y3 , σ z3 ). Through the Bayesian fusion algorithm, the fused position information (x, y, z) and the standard deviation (σ x , σ y , σ z ) are obtained.

[0053] For the x coordinate:

[0054] , ,

[0055] K = + +

[0056]

[0057]

[0058] The same applies to the y coordinate and the z coordinate. Among the values of the reciprocals of the standard deviations , , , the smaller the standard deviation of the sensor (the more reliable the measurement result), the larger its reciprocal, and the greater the weight it occupies in the fusion calculation. In this way, the measurement accuracies of different sensors can be comprehensively considered, higher weights can be assigned to more accurate measurement results, and a more accurate and reliable position estimation of pedestrians and non-motor vehicles can be obtained.

[0059] Since the performances of different sensors vary under different environmental conditions (such as lighting, weather, occlusion, etc.), the accuracy of measuring the position of pedestrians will also vary. The use of the standard deviation can quantify this difference and incorporate it into the calculation process of information fusion, making the finally fused pedestrian position information more accurate and reliable, thereby improving the accuracy and reliability of the entire system for detecting pedestrians in the blind area behind elevated columns. By considering the standard deviation in the system, the characteristics and advantages of different sensors are fully utilized, and their measurement uncertainties are taken into account, avoiding measurement errors caused by the limitations of a single sensor, and providing a statistical and mathematical basis for the performance improvement of the system and more accurate pedestrian position judgment.

[0060] The speed calculation method of the target to be measured includes:

[0061] Obtain the fused position information of the target to be measured at time t1 (x t1 , y t1 , z t1 ), and the fused position information at time t2 (x t2 , y t2 , z t2 ), Δt = t2 - t1.

[0062] The speed of the target to be measured :

[0063]

[0064] The acceleration calculation method includes:

[0065] Obtain the fused position information of the target to be measured at time t3 (x t3 , y t3 , z t3 ).

[0066] The acceleration of the target to be measured in the x direction :

[0067]

[0068] The accelerations of the target to be measured in the y and z directions , Similarly.

[0069] Furthermore, the method for comprehensively evaluating the danger level includes:

[0070]

[0071] d is the vertical distance from the fused position information to the edge of the vertical column blind area close to the side of the second road ( Figure 1 the motor vehicle lane in it).

[0072] By comprehensively considering the position coordinates of the pedestrian, the distance d, speed, and acceleration factors, the danger level is calculated and quantified, which enables the system to more accurately evaluate the dangerous situation of the pedestrian and avoids the ambiguity and uncertainty of subjective judgment. The system can reasonably allocate resources according to different danger levels. For example, for a higher danger level, the system can increase the scanning frequency of the sensor or increase the volume and intensity of the alarm to ensure that the driver can react in time.

[0073] Decision-making steps: According to the speed, acceleration, and distance from the target to be measured to the second road, comprehensively evaluate the danger level, and issue different warnings to the vehicles on the second road under different danger levels.

[0074] As Figure 3 shown, when there is no target to be measured in the blind area of the column, D = 0, which is the first degree of danger. A green signal lamp "Pay attention to the blind area, slow down" is used to give a warning to the vehicles on the second road. When D belongs to (0, D0), it is the second degree of danger, and a red signal lamp "Pay attention to non-motor vehicles breaking in" and a yellow signal lamp (attention sign) are switched to give a warning to the vehicles on the second road. When D is greater than or equal to D0, it is the third degree of danger, and a red signal lamp "Pay attention to non-motor vehicles breaking in" and a projection lamp project "Pay attention to non-motor vehicles breaking in" onto the road surface of the second road to give a warning to the vehicles on the second road. Through several months of testing, according to the data provided by the traffic management department, the present invention can reduce the accident rate by more than 50% year-on-year.

[0075] The present invention also provides a warning system for traffic accidents in the blind area of the column. The warning system for traffic accidents in the blind area of the column can be realized by executing the process steps of the warning method for traffic accidents in the blind area of the column. That is, those skilled in the art can understand the warning method for traffic accidents in the blind area of the column as a preferred implementation manner of the warning system for traffic accidents in the blind area of the column. The system includes:

[0076] Camera: When the first road where the blind area of the column is located is a red light, collect the video of the blind area of the column;

[0077] Target detection model: Obtain the video of the blind area of the column collected by the camera, and detect whether there is a target to be measured in the blind area of the column;

[0078] Processing module: Build a coordinate system for the blind area of the column in the video. When a target to be measured is detected in the blind area of the column, locate the target to be measured through the video and the sensor respectively, calculate the speed and acceleration of the target to be measured, and comprehensively evaluate the degree of danger according to the speed, acceleration of the target to be measured and the distance from the second road. Under different degrees of danger, give different warnings to the vehicles on the second road.

[0079] Those skilled in the art know that in addition to implementing the system and its various devices, modules, and units provided by the present invention in the form of pure computer-readable program code, the method steps can be logically programmed to make the system and its various devices, modules, and units provided by the present invention be implemented in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers to achieve the same function. Therefore, the system and its various devices, modules, and units provided by the present invention can be regarded as a hardware component, and the devices, modules, and units included therein for implementing various functions can also be regarded as the structure within the hardware component; the devices, modules, and units for implementing various functions can also be regarded as both software modules for implementing the method and the structure within the hardware component.

[0080] In the description of the present application, it should be understood that the orientation or positional relationship indicated by the terms "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present application and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation to the present application.

[0081] The specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the above specific embodiments, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Without conflict, the embodiments of the present application and the features in the embodiments can be combined with each other arbitrarily.

Claims

1. A method for warning traffic accidents in pillar blind spots, characterized in that: include: Target recognition step: when the traffic light of the first road where the pillar blind spot is located is red, the video of the pillar blind spot is collected by the camera, and the video is input into the target detection model to detect whether there is a target to be detected in the pillar blind spot; Motion state calculation steps: construct a coordinate system for the blind area of ​​the column in the video. When a target to be measured is detected in the blind area of ​​the column, the target to be measured is located through the video and the sensor, and the speed and acceleration of the target to be measured are calculated; Decision-making steps: Based on the speed, acceleration and distance of the target to be detected from the second road, a comprehensive assessment is made to obtain the degree of danger. Different warnings are issued to vehicles on the second road at different degrees of danger. The methods for locating the target to be measured include: According to the position of the target to be measured in the video, the first position information (x1, y1, z1) of the target to be measured is determined, and the standard deviation of its position is (σ x1 , σ y1 , σ z1 ), the second position information (x2, y2, z2) of the target to be measured is determined by the millimeter wave radar, and the standard deviation of its position is (σ x2 , σ y2 , σ z2 ), the third position information (x3, y3, z3) of the target to be measured is determined by the laser sensor, and the standard deviation of the position is (σ x3 , σ y3 , σ z3 ), through the Bayesian fusion algorithm, the fused position information (x, y, z) and standard deviation (σ x , σ y , σ z ); For the x coordinate: ， ， K= + + The same goes for the y-coordinate and the z-coordinate; The speed calculation methods of the target to be measured include: Get the fusion position information of the target to be measured at time t1 (x t1 ,y t1 , z t1 ), and the fused position information at time t2 (x t2 ,y t2 , z t2 ), Δt= t2- t1; The speed of the target to be measured : ； The calculation method of the acceleration of the target to be measured includes: Get the fusion position information of the target to be measured at time t3 (x t3 ,y t3 , z t3 ) The acceleration of the target to be measured in the x direction : The acceleration of the target to be measured in the y and z directions , Similarly; Comprehensive assessment of the degree of risk includes: d is the vertical distance between the fused position information and the edge of the column blind spot close to the second road side.

2. The pillar blind spot traffic accident warning method according to claim 1, characterized in that: When there is no target to be measured in the blind spot of the column, D=0, which is the first danger level, and the vehicles on the second road are warned by the green signal light. When D belongs to (0, D0), it is the second danger level, and the vehicles on the second road are warned by switching between red and yellow signal lights. When D is greater than or equal to D0, it is the third danger level, and the vehicles on the second road are warned by the red signal light and the projection light.

3. A pillar blind spot traffic accident warning system, characterized in that: The pillar blind spot traffic accident warning system implements the pillar blind spot traffic accident warning method according to claim 1, and the pillar blind spot traffic accident warning system includes: Camera: When the traffic light on the first road where the pillar blind spot is located is red, collect the video of the pillar blind spot; Target detection model: obtain the video of the blind spot of the column collected by the camera, and detect whether there is a target to be detected in the blind spot of the column; Processing module: Construct a coordinate system for the blind spot of the pillar in the video. When a target to be measured is detected in the blind spot of the pillar, the target to be measured is located through video and sensors respectively, and the speed and acceleration of the target to be measured are calculated. According to the speed and acceleration of the target to be measured and the distance from the second road, a comprehensive assessment is made to obtain the degree of danger. Under different degrees of danger, different warnings are issued to vehicles on the second road.

4. The pillar blind spot traffic accident warning system according to claim 3, characterized in that: The sensors include: millimeter wave radar and laser sensor.

5. The pillar blind spot traffic accident warning system according to claim 3, characterized in that: It also includes a signal light and a projection light, through which an early warning is given.

Images