A dual-module ranging system of camera and laser radar

Abstract

The present invention discloses a dual-module ranging system of a camera and a laser radar, which relates to the technical field of vehicle distance detection and improves the accuracy of ranging results in extreme environments. The present invention divides a channel area to be processed from lateral driving image data and central driving image data, performs masking and removal operations on the channel area to be processed, establishes a vehicle periphery model and a traffic road image model according to the lateral driving image data and the central driving image data, sets a number of ranging time nodes, predicts the estimated relative vehicle distance at the next ranging time node according to the relative speed and relative vehicle distance of each driving vehicle, and determines whether the estimated relative vehicle distance is correct according to the real-time driving speed until the corresponding driving vehicle stops driving.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle distance detection, and in particular to a camera and laser radar dual-module distance measurement system. Background Art

[0002] With the development of technologies such as autonomous driving, drone navigation, and intelligent robots, the importance of environmental perception systems in these fields has become increasingly prominent. Cameras and LiDAR, as two main sensors, are widely used in target detection and ranging. Cameras have the advantages of high resolution, rich color information, and low cost, and can capture detailed image information of the environment, but they have certain limitations in distance measurement. In contrast, LiDAR can directly obtain accurate distance information by emitting laser pulses and measuring the return time. It has the advantages of high precision and long-distance measurement, but its cost is relatively high.

[0003] The existing camera and lidar ranging technologies have the following defects:

[0004] Poor environmental adaptability: In complex environments (such as rainy and snowy weather, strong light reflection, etc.), the ranging performance of existing systems will be significantly reduced, and it is difficult to ensure stability and reliability under various environmental conditions.

[0005] High calibration complexity: The spatial coordinate systems of cameras and lidar are different, and a complex calibration process is required to achieve accurate sensor fusion. Existing calibration methods usually rely on manually set calibration objects, which are complex to operate and easily affected by the environment, making it difficult to achieve automated and high-precision calibration.

[0006] Therefore, how to improve the accuracy of the ranging results in extreme environments while improving the data fusion accuracy of the camera and the lidar is a difficulty in the existing technology. For this purpose, a camera and lidar dual-module ranging system is provided. Summary of the invention

[0007] In order to solve the above technical problems, the purpose of the present invention is to provide a camera and laser radar dual-module ranging system.

[0008] In order to achieve the above object, the present invention provides the following technical solutions:

[0009] A camera and laser radar dual-module distance measurement system, comprising a cloud computing platform, wherein the cloud computing platform is communicatively connected with a road data acquisition module, a road model building module and a vehicle distance detection module;

[0010] The road data collection module is provided with an edge data collection unit and a vehicle data collection unit;

[0011] The edge data collection unit is provided with a plurality of roadside units on both sides of the traffic road. The roadside units are provided with high-speed cameras and laser radar devices for collecting the sideways driving image data and the sideways vehicle laser spectrum of each vehicle on the traffic road;

[0012] The vehicle data acquisition unit is used to collect central driving image data and central vehicle laser spectrum with the vehicle in which the two are located as the center through a high-speed camera and a laser radar device;

[0013] The road model building module is used to divide the channel area to be processed from the lateral driving image data and the central driving image data, and after performing a masking and removing operation on the channel area to be processed, a local vehicle periphery model and a local road image model are established according to the lateral driving image data and the central driving image data, and the local vehicle periphery model and the local road image model are matched and overlapped to obtain a vehicle periphery model and a traffic road image model;

[0014] The vehicle distance detection module is used to set a number of ranging time nodes, and obtain the relative speed between the moving vehicles according to the central vehicle laser spectrum, and obtain the real-time driving speed and relative vehicle distance of each moving vehicle according to the side vehicle laser spectrum, and then predict the estimated relative vehicle distance at the next ranging time node according to the relative speed and relative vehicle distance of each moving vehicle, and judge whether the estimated relative vehicle distance is correct according to the real-time driving speed until the corresponding moving vehicle stops.

[0015] Furthermore, the collection process of the central driving image data, the central vehicle laser spectrum, the side driving image data and the side vehicle laser spectrum includes:

[0016] Several roadside units are installed on both sides of the traffic road, each roadside unit is connected to the edge data collection unit for communication, and the same number is set for the roadside units in the same row;

[0017] The roadside unit is equipped with a high-speed camera and a laser radar device, wherein the data collection range of the high-speed camera and the laser radar device is the same, and the data collection ranges of the high-speed cameras and the laser radar devices of the roadside units adjacent to each other in the same column and in the same row overlap;

[0018] When a moving vehicle passes the first roadside unit, the vehicle data collection unit associated with the moving vehicle communicates with the edge data collection unit, and then the edge data collection unit sets a unique identity number for the corresponding vehicle data collection unit;

[0019] When a vehicle is traveling on a traffic road, each set of high-speed cameras and lidar devices located on the vehicle, as well as the high-speed cameras and lidar devices on each roadside unit, continuously collect central driving image data, central vehicle laser spectrum, side driving image data and side vehicle laser spectrum within their data collection range.

[0020] Furthermore, the process of dividing the channel area to be processed includes:

[0021] Obtain the RGB channel value of each pixel in the central driving image data and the side driving image data, and set the channel threshold;

[0022] Obtain the mask channel of each pixel in each center driving image data and the side driving image data, compare the mask channel value of each pixel with the channel threshold, mark the corresponding pixel as a pixel to be processed according to the comparison result, connect each pixel to be processed in sequence, and mark the channel area to be processed on the corresponding image data according to the connection result.

[0023] Furthermore, the process of performing a mask removal operation on the channel area to be processed includes:

[0024] The maximum mask channel value in the channel area to be processed is selected and recorded as the atmospheric light correction parameter, and the transmittance of the corresponding pixel to be processed is obtained according to the mask channel value, and then the mask removal operation is performed on each pixel to be processed in each channel area to be processed according to the atmospheric light correction parameter and the transmittance;

[0025] After all the central driving image data and the side driving image data have been subjected to the mask removal operation, the corresponding local vehicle periphery model and the local road image model are generated according to each of the central driving image data and the side driving image data.

[0026] Furthermore, the process of establishing the vehicle periphery model and the traffic road image model includes:

[0027] A plurality of matching model regions are randomly set in the local vehicle periphery model and the local road image model, and the matching model regions in the local vehicle periphery model or the local road image model with the same identity number or roadside unit number are matched and overlapped with each other;

[0028] A new matching model area is divided again according to the matching result until there is no matching matching model area between the two local vehicle periphery models or the local road image models;

[0029] According to the model of the matching model area between any two local vehicle periphery models or local road image models, the corresponding local vehicle periphery models are overlapped and spliced ​​to obtain the vehicle periphery model and the road image model. According to the roadside unit number corresponding to each road image model, each road image model is spliced ​​in sequence to obtain the traffic road image model.

[0030] Furthermore, the process of acquiring the relative speed of the traveling vehicles and the real-time traveling speed between the traveling vehicles includes:

[0031] Each moving vehicle in the vehicle peripheral model is recorded as a moving associated vehicle, and according to the position of each moving associated vehicle in the vehicle peripheral model, the corresponding central vehicle laser spectrum is retrieved;

[0032] Obtaining the initial frequency of the laser signal emitted by the laser radar device on the moving vehicle, and obtaining the reflection spectrum of the laser spectrum of the center vehicle, thereby obtaining the relative speed of each moving associated vehicle relative to the moving vehicle corresponding to the vehicle peripheral model;

[0033] According to the lateral vehicle laser spectrum generated by each roadside unit, the real-time driving speed of each vehicle within the data collection range at the current moment is obtained.

[0034] Furthermore, the process of obtaining the estimated relative vehicle distance includes:

[0035] A number of ranging time nodes are set, and the vehicle peripheral model is mapped into the traffic road image model, and then a number of driving vehicle models are planned in the traffic road image model, and the relative inclination angles between the driving vehicle models in the traffic road image model are obtained. At the same time, the relative vehicle distances between the adjacent driving vehicle models are obtained according to the central vehicle laser spectrum of the adjacent driving vehicle models at the ranging time node, and then the relative vehicle distances between the driving vehicle models on the traffic road image model are obtained in sequence;

[0036] The real-time driving speed of each driving vehicle model at the current ranging time node and the relative speed between each driving vehicle are marked on the driving vehicle model, and then according to the relative vehicle distance L and relative speed between each driving vehicle model at the current ranging time node, the expected relative vehicle distance between each driving vehicle at the next ranging time node is predicted.

[0037] Furthermore, the process of judging whether the estimated relative vehicle distance is correct according to the real-time driving speed includes:

[0038] At the same time, the estimated arrival position of each driving vehicle is determined according to the current driving speed v of each driving vehicle at the previous ranging time node, and then the lateral driving image data and the lateral vehicle laser spectrum generated by the roadside unit at the expected arrival position are retrieved at the next ranging time node, and the actual relative vehicle distance of each driving vehicle at the next ranging time node is obtained according to the lateral driving image data and the lateral vehicle laser spectrum, and the actual relative vehicle distance is compared with the estimated relative vehicle distance. If the difference between the two is between 0.95, no operation is performed, otherwise the estimated relative vehicle distance between the corresponding driving vehicles is recalculated at half the time of the next ranging time node;

[0039] The above process of obtaining the estimated relative vehicle distance and the actual relative vehicle distance is repeated until the moving vehicle stops.

[0040] Compared with the prior art, the present invention has the following beneficial effects:

[0041] The present invention divides a to-be-processed channel area from the lateral driving image data and the central driving image data, performs masking and removing operations on the to-be-processed channel area, establishes a vehicle periphery model and a traffic road image model according to the lateral driving image data and the central driving image data, sets a plurality of ranging time nodes, obtains the relative speed between the driving vehicles according to the central vehicle laser spectrum, and obtains the real-time driving speed and relative vehicle distance of each driving vehicle according to the lateral vehicle laser spectrum, and then predicts the estimated relative vehicle distance at the next ranging time node according to the relative speed and relative vehicle distance of each driving vehicle, and judges whether the estimated relative vehicle distance is correct according to the real-time driving speed until the corresponding driving vehicle stops driving, thereby improving the accuracy of the data fusion of the camera and the laser radar while improving the accuracy of the ranging result in extreme environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 It is a schematic diagram of the present invention. DETAILED DESCRIPTION

[0043] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation mode, structure, characteristics and effects of the present invention are described in detail below in combination with the accompanying drawings and preferred embodiments.

[0044] like Figure 1 As shown, a camera and laser radar dual-module distance measurement system includes a cloud computing platform, and the cloud computing platform is communicatively connected with a road data acquisition module, a road model building module, and a vehicle distance detection module;

[0045] The road data collection module is provided with an edge data collection unit and a vehicle data collection unit;

[0046] The edge data collection unit is provided with a plurality of roadside units on both sides of the traffic road. The roadside units are provided with high-speed cameras and laser radar devices for collecting the sideways driving image data and the sideways vehicle laser spectrum of each vehicle on the traffic road;

[0047] The vehicle data acquisition unit is used to collect central driving image data and central vehicle laser spectrum with the vehicle in which the two are located as the center through a high-speed camera and a laser radar device;

[0048] The road model building module is used to divide the channel area to be processed from the lateral driving image data and the central driving image data, and after performing a masking and removing operation on the channel area to be processed, a local vehicle periphery model and a local road image model are established according to the lateral driving image data and the central driving image data, and the local vehicle periphery model and the local road image model are matched and overlapped to obtain a vehicle periphery model and a traffic road image model;

[0049] The vehicle distance detection module is used to set a number of ranging time nodes, divide a plurality of driving-related vehicles in the vehicle peripheral model, and obtain the relative speed between the driving vehicles according to the central vehicle laser spectrum, and obtain the real-time driving speed and relative vehicle distance of each driving vehicle according to the side vehicle laser spectrum, and then predict the estimated relative vehicle distance at the next ranging time node according to the relative speed and relative vehicle distance of each driving vehicle, and judge whether the estimated relative vehicle distance is correct according to the real-time driving speed until the corresponding driving vehicle stops.

[0050] Further, the working principle of the present invention is specifically described below through embodiments:

[0051] 2n roadside units are installed on both sides of the traffic road, and each roadside unit is connected to the edge data collection unit for communication. Then, the edge data collection unit sets the same number for the roadside units in the same row, where the numbers can be s1, s2, s3, ..., s n , where n is a natural number greater than 0;

[0052] The roadside unit is equipped with a high-speed camera and a laser radar device, wherein the data collection range of the high-speed camera and the laser radar device is the same, and the data collection ranges of the high-speed cameras and the laser radar devices of the roadside units adjacent to each other in the same column and in the same row overlap;

[0053] The moving vehicle is provided with six groups of high-speed cameras and laser radar devices, which are respectively located at the front, door and both sides of the rear of the moving vehicle, and the data collection ranges of each group of high-speed cameras and laser radar devices are the same and cross-sectional, so that the combined result of the data collection ranges of each group of high-speed cameras and laser radar devices covers a circle of the moving vehicle, and at the same time, each moving vehicle is communicatively associated with a vehicle data collection unit;

[0054] When a moving vehicle passes the first roadside unit, the vehicle data collection unit associated with the moving vehicle communicates with the edge data collection unit, and then the edge data collection unit sets a unique identity number for the corresponding vehicle data collection unit;

[0055] When the vehicle is traveling on the traffic road, each set of high-speed cameras and laser radar devices located on the vehicle, as well as the high-speed cameras and laser radar devices on each roadside unit, continuously collects the center driving image data, the center vehicle laser spectrum, the side driving image data and the side vehicle laser spectrum within their data collection range;

[0056] It should be noted that the frequencies of the laser signals emitted by the roadside unit and the laser radar device on the moving vehicle are different.

[0057] Furthermore, the road data acquisition module marks the newly generated central driving image data, central vehicle laser spectrum, side driving image data and side vehicle laser spectrum with corresponding identity numbers or roadside unit numbers, and then synchronously sends all the data to the road model building module and the vehicle distance detection module;

[0058] The road model building module first obtains the RGB channel value of each pixel in the central driving image data and the side driving image data, and sets the channel threshold;

[0059] Then, the mask channel of each pixel in each central driving image data and the side driving image data is obtained, wherein the selection formula of the mask channel is: ,in Represent the RGB channel values ​​of the pixel respectively, J(x) represents the mask channel value of the x-th pixel, and x is a natural number greater than 0;

[0060] Compare the mask channel value of each pixel with the channel threshold. If the mask channel value is less than or equal to the channel threshold, ignore the corresponding pixel. If the mask channel value is greater than the channel threshold, mark the corresponding pixel as a pixel to be processed.

[0061] Connect each pixel to be processed in sequence, mark the channel area to be processed on the corresponding image data according to the connection result, select the maximum mask channel value in the channel area to be processed as the atmospheric light correction parameter θ, and obtain the transmittance h(x) of the corresponding pixel to be processed according to the mask channel value, where the calculation formula of h(x) is: ,in Represents the weight parameter, which usually takes a value between 0.95 and 1.0;

[0062] Then, a mask removal operation is performed on each pixel to be processed in each channel area to be processed, wherein the mask removal operation formula is: , where I(x) represents the value of the mask channel value of the corresponding pixel to be processed after the mask removal operation.

[0063] Furthermore, after all the central driving image data and the side driving image data have been subjected to the mask removal operation, a corresponding local vehicle periphery model and a local road image model are generated according to each of the central driving image data and the side driving image data;

[0064] Since the central driving image data and the side driving image data with the same identification number or roadside unit number are collected based on the same driving vehicle or the same section of traffic road, there must be the same parts between the local vehicle peripheral models or local road image models generated corresponding to the central driving image data or the side driving image data with the same identification number or roadside unit number;

[0065] Therefore, a number of matching model regions are randomly set inside the local vehicle periphery model and the local road image model, and the matching model regions in the local vehicle periphery model or the local road image model with the same identity number or roadside unit number are matched and overlapped with each other;

[0066] If the matching model regions from different local vehicle periphery models or local road image models are not the same, it is determined that the corresponding two matching model regions do not match, and the next matching model region matching operation is performed;

[0067] If the matching model areas from different local vehicle periphery models or local road image models are the same, eight matching model areas of the same size are divided around the corresponding matching model areas, and then the newly divided matching model areas are matched, and new matching model areas are divided again according to the matching results, until there is no matching matching model area between the two local vehicle periphery models or local road image models;

[0068] According to the models of the matching model areas between any two local vehicle periphery models or local road image models, the corresponding local vehicle periphery models are overlapped and spliced ​​to obtain the vehicle periphery model and the road image model;

[0069] The process of generating the road image model is adopted, and the road image models are sequentially spliced ​​together according to the roadside unit numbers corresponding to the road image models to obtain the traffic road image model.

[0070] Furthermore, the road model building module synchronously sends the vehicle periphery model and traffic road image model generated at each moment to the vehicle distance detection module;

[0071] Each moving vehicle in the vehicle peripheral model is recorded as a moving associated vehicle, and according to the position of each moving associated vehicle in the vehicle peripheral model, the corresponding central vehicle laser spectrum is retrieved;

[0072] Obtain the initial frequency f0 of the laser signal emitted by the laser radar device on the moving vehicle, and obtain the reflection spectrum f of the laser spectrum of the central vehicle, and then obtain the relative speed v of each moving associated vehicle relative to the corresponding moving vehicle of the vehicle peripheral model 相对 , where the relative speed v 相对 The calculation formula is: , where c is the speed of light;

[0073] Since the speed of each roadside unit is constant at 0 at each moment, the relative speed between each moving vehicle and the roadside unit is the real-time speed v of each moving vehicle. Then, according to the side vehicle laser spectrum generated by each roadside unit, the real-time speed v of each moving vehicle within the data collection range at the current moment is obtained;

[0074] A number of ranging time nodes are set, and the vehicle peripheral model is mapped into the traffic road image model, and then a number of driving vehicle models are planned in the traffic road image model, and the relative inclination angle β between each driving vehicle model in the traffic road image model is obtained. At the same time, the relative vehicle distance between each adjacent driving vehicle model is obtained according to the central vehicle laser spectrum of the adjacent driving vehicle model at the ranging time node, and then the relative vehicle distance L between each driving vehicle model on the traffic road image model is obtained in turn;

[0075] The real-time driving speed v of each driving vehicle model at the current ranging time node, as well as the relative speed v between each driving vehicle 相对 Marked on the driving vehicle model, and then according to the relative distance L and relative speed v between each driving vehicle model at the current distance measurement time node 相对 , predict the estimated relative distance L between each vehicle at the next distance measurement time node 预计, where the estimated relative vehicle distance L 预计 The calculation formula is: , where time represents the time interval between two adjacent ranging time nodes, and the relative tilt angle β ranges from 0 to 90 degrees;

[0076] At the same time, the estimated arrival position of each driving vehicle is determined according to the current driving speed v of each driving vehicle at the previous ranging time node, and then the lateral driving image data and lateral vehicle laser spectrum generated by the roadside unit at the expected arrival position are retrieved at the next ranging time node. The actual relative vehicle distance of each driving vehicle at the next ranging time node is obtained according to the lateral driving image data and the lateral vehicle laser spectrum, and the actual relative vehicle distance is compared with the estimated relative vehicle distance L 预计 If the difference between the two is within 0.95, no operation is performed. Otherwise, the estimated relative distance L between the corresponding vehicles is recalculated at half the time of the next distance measurement time node. 预计 ;

[0077] Repeat the above to obtain the estimated relative vehicle distance L 预计 and the actual relative vehicle distance until the moving vehicle stops.

[0078] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technical personnel in this field can make some changes or modify the technical contents disclosed above into equivalent embodiments without departing from the scope of the technical solution of the present invention. However, any brief modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. A camera and laser radar dual-module ranging system, including a cloud computing platform, characterized in that: The cloud computing platform is communicatively connected with a road data acquisition module, a road model building module and a vehicle distance detection module; The road data collection module is provided with an edge data collection unit and a vehicle data collection unit; The edge data collection unit is provided with a plurality of roadside units on both sides of the traffic road. The roadside units are provided with high-speed cameras and laser radar devices for collecting the sideways driving image data and the sideways vehicle laser spectrum of each vehicle on the traffic road; The vehicle data acquisition unit is used to collect central driving image data and central vehicle laser spectrum with the vehicle in which the two are located as the center through a high-speed camera and a laser radar device; The road model building module is used to divide the channel area to be processed from the lateral driving image data and the central driving image data, and after performing a masking and removing operation on the channel area to be processed, a local vehicle periphery model and a local road image model are established according to the lateral driving image data and the central driving image data, and the local vehicle periphery model and the local road image model are matched and overlapped to obtain a vehicle periphery model and a traffic road image model; The vehicle distance detection module is used to set a number of ranging time nodes, and obtain the relative speed between the moving vehicles according to the central vehicle laser spectrum, and obtain the real-time driving speed and relative vehicle distance of each moving vehicle according to the side vehicle laser spectrum, and then predict the estimated relative vehicle distance at the next ranging time node according to the relative speed and relative vehicle distance of each moving vehicle, and judge whether the estimated relative vehicle distance is correct according to the real-time driving speed, until the corresponding moving vehicle stops moving; The process of obtaining the estimated relative vehicle distance includes: A number of ranging time nodes are set, and the vehicle peripheral model is mapped into the traffic road image model, and then a number of driving vehicle models are planned in the traffic road image model, and the relative inclination angles between the driving vehicle models in the traffic road image model are obtained. At the same time, the relative vehicle distances between the adjacent driving vehicle models are obtained according to the central vehicle laser spectrum of the adjacent driving vehicle models at the ranging time node, and then the relative vehicle distances between the driving vehicle models on the traffic road image model are obtained in sequence; The real-time driving speed of each driving vehicle model at the current ranging time node and the relative speed between each driving vehicle are marked on the driving vehicle model, and then according to the relative vehicle distance and relative speed between each driving vehicle model at the current ranging time node, the expected relative vehicle distance between each driving vehicle at the next ranging time node is predicted.

2. A camera and laser radar dual-module ranging system according to claim 1, characterized in that: The collection process of the central driving image data, the central vehicle laser spectrum, the side driving image data and the side vehicle laser spectrum includes: Several roadside units are installed on both sides of the traffic road, each roadside unit is connected to the edge data collection unit for communication, and the same number is set for the roadside units in the same row; When a moving vehicle passes the first roadside unit, the vehicle data collection unit associated with the moving vehicle communicates with the edge data collection unit, and then the edge data collection unit sets a unique identity number for the corresponding vehicle data collection unit; When a vehicle is traveling on a traffic road, each set of high-speed cameras and lidar devices located on the vehicle, as well as the high-speed cameras and lidar devices on each roadside unit, continuously collect central driving image data, central vehicle laser spectrum, side driving image data and side vehicle laser spectrum within their data collection range.

3. A camera and laser radar dual-module ranging system according to claim 2, characterized in that: The process of dividing the channel area to be processed includes: The RGB channel value of each pixel in the central driving image data and the lateral driving image data is obtained, and the channel threshold is set, and the mask channel of each pixel in each central driving image data and the lateral driving image data is obtained, and the mask channel value of each pixel is compared with the channel threshold, and the corresponding pixel is marked as a pixel to be processed according to the comparison result, and each pixel to be processed is connected in sequence, and the channel area to be processed is marked on the corresponding image data according to the connection result.

4. A camera and laser radar dual-module ranging system according to claim 3, characterized in that: The process of performing mask removal operation on the channel area to be processed includes: The maximum mask channel value in the channel area to be processed is selected and recorded as the atmospheric light correction parameter, and the transmittance of the corresponding pixel to be processed is obtained according to the mask channel value, and then the mask removal operation is performed on each pixel to be processed in each channel area to be processed according to the atmospheric light correction parameter and the transmittance; After all the central driving image data and the side driving image data have been subjected to the mask removal operation, the corresponding local vehicle periphery model and the local road image model are generated according to each of the central driving image data and the side driving image data.

5. A camera and laser radar dual-module ranging system according to claim 4, characterized in that: The process of establishing the vehicle periphery model and the traffic road image model includes: A plurality of matching model regions are randomly set in the local vehicle periphery model and the local road image model, and the matching model regions in the local vehicle periphery model or the local road image model with the same identity number or roadside unit number are matched and overlapped with each other; A new matching model area is divided again according to the matching result until there is no matching matching model area between the two local vehicle periphery models or the local road image models; According to the model of the matching model area between any two local vehicle periphery models or local road image models, the corresponding local vehicle periphery models are overlapped and spliced ​​to obtain the vehicle periphery model and the traffic road image model.

6. A camera and laser radar dual-module ranging system according to claim 5, characterized in that: The process of acquiring the relative speed of the traveling vehicles and the real-time traveling speed between the traveling vehicles includes: Each moving vehicle in the vehicle peripheral model is recorded as a moving associated vehicle, and according to the position of each moving associated vehicle in the vehicle peripheral model, the corresponding central vehicle laser spectrum is retrieved; Obtaining the initial frequency of the laser signal emitted by the laser radar device on the moving vehicle, and obtaining the reflection spectrum of the laser spectrum of the center vehicle, thereby obtaining the relative speed of each moving associated vehicle relative to the moving vehicle corresponding to the vehicle peripheral model; According to the lateral vehicle laser spectrum generated by each roadside unit, the real-time driving speed of each vehicle within the data collection range at the current moment is obtained.

7. A camera and laser radar dual-module ranging system according to claim 6, characterized in that: The process of judging whether the estimated relative vehicle distance is correct based on the real-time driving speed includes: At the same time, the estimated arrival position of each driving vehicle is determined according to the current driving speed of each driving vehicle at the previous ranging time node, and then the lateral driving image data and the lateral vehicle laser spectrum generated by the roadside unit of the estimated arrival position are retrieved at the next ranging time node, and the actual relative vehicle distance of each driving vehicle at the next ranging time node is obtained according to the lateral driving image data and the lateral vehicle laser spectrum, and the actual relative vehicle distance is compared with the estimated relative vehicle distance, and the estimated relative vehicle distance between the corresponding driving vehicles is recalculated according to the comparison result at half the time of the next ranging time node, and the process of obtaining the estimated relative vehicle distance and the actual relative vehicle distance is repeated until the driving vehicle stops.

Images