Radar camera calibration quality evaluation method and system
By using error calculation and cross-verification of pixel coordinate system and world coordinate system in radar camera calibration, combined with the Local Outlier Factor method, the problem of too long calibration time and too much labor is solved, achieving more efficient calibration accuracy and intuitiveness software display.
Patent Information
- Application Number
- CN202111516455.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-12-09
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2041-12-09
AI Technical Summary
The calibration process of existing radar cameras is too long, too much manpower is invested, and the accuracy is not high, making it difficult to complete the calibration process quickly and accurately.
By performing error calculation and cross-verification on pixel coordinate systems and world coordinate systems, outliers are detected using the Local Outlier Factor method, and the deviation points are quickly positioned and corrected through the aerial view of the map to achieve accurate mapping and calibration of radar, camera and GPS coordinates .
It significantly shortens the calibration time, improves the perceived distance accuracy of the road measurement perception scheme, improves the accuracy of the accuracy of 35%, and makes the calibration process more intuitive and convenient.
Smart Images

Figure CN114217277B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of radar photography, and in particular to a radar camera calibration quality evaluation method and system. Background Art
[0002] Calibration mainly refers to the use of standard measuring instruments to test the accuracy (precision) of the instrument in use to see if it meets the standard. It is generally used for instruments with higher precision. Calibration can also be considered as calibration. Therefore, it can also be considered that calibration includes the above two meanings.
[0003] A patent document with publication number CN112684424A discloses an automatic calibration method for a millimeter-wave radar and a camera, comprising the following steps: 1) setting and only setting one moving target within the cross-visible range of the millimeter-wave radar and the camera, and the moving target moves to multiple positions within the cross-visible range; 2) the millimeter-wave radar and the camera are sampled at the same starting time and the same frequency; 3) a preliminary screening is performed on multiple target points of the millimeter-wave radar and the camera in each sampling period; 4) the target points obtained by the millimeter-wave radar and the camera for multiple consecutive sampling periods are obtained respectively, and outlier screening is performed; 5) the radar target point coordinates and the camera target point coordinates of the consecutive sampling periods after the outlier screening are obtained, and a neural network is constructed for training, and automatic calibration of the millimeter-wave radar and the camera is realized according to the trained neural network.
[0004] With respect to the above-mentioned related technologies, the inventor believes that the above-mentioned technologies have the problems of too long calibration time and too much manpower investment. Therefore, it is necessary to propose a technical solution to improve the above-mentioned technical problems. Summary of the invention
[0005] In view of the defects in the prior art, an object of the present invention is to provide a radar camera calibration quality evaluation method and system.
[0006] According to a radar camera calibration quality evaluation method provided by the present invention, the method comprises the following steps:
[0007] Step S1: Calculate the error of the pixel coordinate system and determine the outliers;
[0008] Step S2: unify pixel coordinate system formula;
[0009] Step S3: unify the world coordinate system and select outliers;
[0010] Step S4: Summarize the pixel domain and world coordinate system results.
[0011] Preferably, step S1 comprises the following steps:
[0012] Step S1.1: Perform difference operations on the positions of the radar, camera and GPS coordinates of N groups of scattered points in pixels and their true values respectively;
[0013] Step S1.2: average the results obtained by the difference operation, sort the N groups of results in descending order, and if the error value obtained is greater than the threshold, it is determined to be an outlier;
[0014] Step S1.3: Use the Local Outlier Factor method to detect outliers in the N groups of values obtained in step S2.1.
[0015] Preferably, in step S3, the true values of pixels, radar and raw data are uniformly mapped to world coordinates, and a map is used for screen display, and outliers are selected through a bird's-eye view.
[0016] Preferably, in step S4, the deviation result of the pixel coordinate system is compared with the result of the bird's-eye view of the map to find the calibration outliers, notify the calibration personnel to collect correction data, and evaluate and deploy the calibration quality of the drive test end.
[0017] Preferably, the radar, camera and calculated GPS world coordinates are mapped to the pixel coordinate system together, and the N target points collected are mapped and calculated to obtain N groups of scattered points in the pixel coordinate system, ensuring that the N groups of scattered points are all located in the ROI road area, and the N groups of points are effectively sorted by errors, the deviation outliers in the scattered points are located, and the user is informed of the results, and the data points are collected again to locate the error points.
[0018] The present invention also provides a radar camera calibration quality evaluation system, the system comprising the following modules:
[0019] Module M1: Calculate the error of the pixel coordinate system and determine the outliers;
[0020] Module M2: Unified pixel coordinate system formula;
[0021] Module M3: Unify the world coordinate system and select outliers;
[0022] Module M4: Aggregate pixel domain and world coordinate system results.
[0023] Preferably, the module M1 includes the following modules:
[0024] Module M1.1: Perform difference operations on the positions of N groups of scattered radar, camera and GPS coordinates in pixels and their true values respectively;
[0025] Module M1.2: average the results obtained by the difference operation, sort the N groups of results in descending order, and if the error value is greater than the threshold, it is determined to be an outlier;
[0026] Module M1.3: Use the Local Outlier Factor system to detect outliers in the N sets of values obtained from module M2.1.
[0027] Preferably, the module M3 uniformly maps the true values of pixels, radar and raw data into world coordinates, uses a map for screen display, and selects outliers through a bird's-eye view.
[0028] Preferably, the module M4 compares the deviation result of the pixel coordinate system with the result of the bird's-eye view of the map to find the calibration outliers, notify the calibration personnel to collect correction data, and evaluate and deploy the calibration quality of the drive test end.
[0029] Preferably, the radar, camera and calculated GPS world coordinates are mapped to the pixel coordinate system together, and the N target points collected are mapped and calculated to obtain N groups of scattered points in the pixel coordinate system, ensuring that the N groups of scattered points are all located in the ROI road area, and the N groups of points are effectively sorted by errors, the deviation outliers in the scattered points are located, and the user is informed of the results, and the data points are collected again to locate the error points.
[0030] Compared with the prior art, the present invention has the following beneficial effects:
[0031] 1. The present invention solves the problem of long calibration time and excessive manpower investment in the RSU road test sensing solution;
[0032] 2. The present invention improves the perception distance accuracy of the combined solution of radar and camera at the road test end by 35% compared with the previous separate calibration solution;
[0033] 3. The present invention supports pixel coordinate system, cross-verification of the accuracy of world coordinate system, and the software display is more intuitive and convenient. BRIEF DESCRIPTION OF THE DRAWINGS
[0034] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:
[0035] Figure 1 It is the process principle diagram of the present invention;
[0036] Figure 2 It is the true value map of radar pixels in the world coordinate system. DETAILED DESCRIPTION
[0037] The present invention is described in detail below in conjunction with specific embodiments. The following embodiments will help those skilled in the art to further understand the present invention, but are not intended to limit the present invention in any form. It should be noted that, for those of ordinary skill in the art, several changes and improvements can also be made without departing from the concept of the present invention. These all belong to the protection scope of the present invention.
[0038] The present invention provides a radar camera calibration quality evaluation method. The radar, camera and calculated GPS world coordinates are mapped to a pixel coordinate system together, and mapping calculation is performed on N collected target points to obtain N groups of scattered points in the pixel coordinate system. It is ensured that the N groups of scattered points are all located in a ROI road area, and then the N groups of points are effectively sorted by errors, and deviation outliers in the scattered points are quickly located. The user is informed of the results, and data points are re-collected to achieve the effect of quickly locating error points.
[0039] A radar camera calibration quality evaluation method comprises the following steps:
[0040] Step 1: Unify the pixel coordinate system error calculation; Step 1.1: Perform difference operations on the positions of the radar, camera, and GPS coordinates of N groups of scattered points in the pixel and the true value respectively; Step 1.2: Then calculate the average and sort the N groups of results in descending order. If the error value obtained is greater than a certain threshold, it is determined to be an outlier; Step 1.3: Use the Local Outlier Factor method to detect the outliers of the N groups of values obtained in step 2.1, and inform the user whether he can apply to use the lof technology for coordinate system error calculation.
[0041] Step 2: Unify the pixel coordinate system formula and corresponding software presentation:
[0042]
[0043] Among them: Rank represents the sorting of N groups of calculation results in brackets;
[0044] DO stands for descending order (sort from largest to smallest);
[0045] Dis stands for finding the Euclidean distance between two points;
[0046] P0-PN represents the coordinate values of the origin of the image after calibration, from group 0 to group N points;
[0047] A0-AN represents the true image value of the calibrated original data;
[0048] R0-RN represents the pixel value of the radar value in the image coordinate system;
[0049] G0-GN represents the pixel value converted from GPS data into the image coordinate system
[0050] Step 3: Unify the world coordinate system; uniformly map pixels, radar, and original data true values to the world coordinate system, and use high-precision maps for screen display. You can quickly select outliers through a bird's-eye view.
[0051] Step 4: Summarize the pixel domain and world coordinate system results; by comparing the deviation results of the pixel coordinate system with the results of the bird's-eye view of the map, it is easy to quickly find the calibration outliers, so as to notify the calibration personnel to quickly collect and correct the data, so as to achieve the calibration quality evaluation and rapid deployment of the road test end.
[0052] The present invention also provides a radar camera calibration quality evaluation system, which includes the following modules: Module M1: Calculate the error of the pixel coordinate system and determine the outliers; Module M1.1: Perform difference operations on the positions of radar, camera and GPS coordinates of N groups of scattered points in pixels and the true values respectively; Module M1.2: Average the results obtained by the difference operations, arrange the N groups of results in descending order, and determine the outliers if the error value obtained is greater than a threshold; Module M1.3: Use the Local Outlier Factor system to detect the outliers using the N groups of values obtained by module M2.1.
[0053] Module M2: Unify the pixel coordinate system formula; Module M3: Unify the world coordinate system and select outliers; Unify the true values of pixels, radar and raw data into world coordinates, use maps for screen display, and select outliers through a bird's-eye view.
[0054] Module M4: Summarize the pixel domain and world coordinate system results. By comparing the deviation results of the pixel coordinate system with the results of the bird's-eye view of the map, find the calibration outliers, notify the calibration personnel to collect correction data, and evaluate and deploy the calibration quality of the road test end.
[0055] The radar, camera and calculated GPS world coordinates are mapped to the pixel coordinate system together, and the N target points collected are mapped and calculated to obtain N groups of scattered points in the pixel coordinate system. It is ensured that the N groups of scattered points are all located in the ROI road area. The N groups of points are effectively sorted by errors, the deviation outliers in the scattered points are located, the user is informed of the results, the data points are re-collected, and the error points are located.
[0056] The present invention solves the problem of long calibration time and excessive manpower investment in the RSU road test sensing solution.
[0057] Those skilled in the art know that, in addition to realizing the system and its various devices, modules, and units provided by the present invention in a purely computer-readable program code, it is entirely possible to realize the same functions in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, the system and its various devices, modules, and units provided by the present invention can be considered as a hardware component, and the devices, modules, and units included therein for realizing various functions can also be regarded as structures within the hardware component; the devices, modules, and units for realizing various functions can also be regarded as both software modules for realizing the method and structures within the hardware component.
[0058] The above describes the specific embodiments of the present invention. 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 does not affect the essence of the present invention. In the absence of conflict, the embodiments of the present application and the features in the embodiments can be combined with each other arbitrarily.
Claims
1. A radar camera calibration quality evaluation method, characterized in that: The method comprises the following steps: Step S1: Calculate the error of the pixel coordinate system and determine the outliers; Step S2: unify pixel coordinate system formula; Step S3: unify the world coordinate system and select outliers; Step S4: Summarize pixel domain and world coordinate system results; The step S1 comprises the following steps: Step S1.1: Perform difference operations on the positions of the radar, camera and GPS coordinates of N groups of scattered points in pixels and their true values respectively; Step S1.2: average the results obtained by the difference operation, sort the N groups of results in descending order, and if the error value obtained is greater than the threshold, it is determined to be an outlier; Step S1.3: Use the Local Outlier Factor method to detect outliers in the N groups of values obtained in step S1.1; In step S3, the true values of pixels, radar and raw data are uniformly mapped to world coordinates, and a map is used for screen display, and outliers are selected through a bird's-eye view; In step S4, the deviation result of the pixel coordinate system is compared with the result of the bird's-eye view of the map to find the calibration outliers, notify the calibration personnel to collect correction data, and evaluate and deploy the calibration quality of the road test end.
2. The radar camera calibration quality evaluation method according to claim 1, characterized in that: The radar, camera and calculated GPS world coordinates are mapped to the pixel coordinate system together, and the N target points collected are mapped and calculated to obtain N groups of scattered points in the pixel coordinate system. It is ensured that the N groups of scattered points are all located in the ROI road area. The N groups of points are effectively sorted by errors, the deviation outliers in the scattered points are located, the user is informed of the results, the data points are re-collected, and the error points are located.
3. A radar camera calibration quality evaluation system, characterized in that: The system includes the following modules: Module M1: Calculate the error of the pixel coordinate system and determine the outliers; Module M2: Unified pixel coordinate system formula; Module M3: Unify the world coordinate system and select outliers; Module M4: Summarize pixel domain and world coordinate system results; The module M1 includes the following modules: Module M1.1: Perform difference operations on the positions of N groups of scattered radar, camera and GPS coordinates in pixels and their true values respectively; Module M1.2: average the results obtained by the difference operation, sort the N groups of results in descending order, and if the error value is greater than the threshold, it is determined to be an outlier; Module M1.3: Use the Local Outlier Factor system to detect outliers in the N sets of values obtained in module M1.1; In the module M3, pixels, radar and original data true values are uniformly mapped to world coordinates, and a map is used for screen display, and outliers are selected through a bird's-eye view; In the module M4, the deviation result of the pixel coordinate system is compared with the result of the bird's-eye view of the map to find the calibration outliers, notify the calibration personnel to collect correction data, and evaluate and deploy the calibration quality of the road test end.
4. The radar camera calibration quality evaluation system according to claim 3, characterized in that: The radar, camera and calculated GPS world coordinates are mapped to the pixel coordinate system together, and the N target points collected are mapped and calculated to obtain N groups of scattered points in the pixel coordinate system. It is ensured that the N groups of scattered points are all located in the ROI road area. The N groups of points are effectively sorted by errors, the deviation outliers in the scattered points are located, the user is informed of the results, the data points are re-collected, and the error points are located.
