Dynamic calibration method and calibration device for vehicle-mounted camera

Abstract

The invention discloses a dynamic calibration method and calibration device for a vehicle-mounted camera, and belongs to the technical field of vehicle-mounted image method calibration. The method comprises the following steps: radar performance data is monitored in real time, radar self-adaptive adjustment is carried out according to rainfall intensity of each monitoring time period, and then a laser signal is emitted; monitoring the signal state data in real time, and judging whether to carry out signal state optimization adjustment or not according to the rainfall intensity of each monitoring time period to obtain an echo signal correction value; processing each received echo signal according to the laser signal and the echo signal correction value to obtain a standard echo signal; according to the invention, quality optimization processing is carried out on the synchronously acquired camera image according to the image quality parameter, and the position of the camera is calibrated according to the standard echo signal and the camera image after quality optimization processing, so that accuracy and stability of calibration of the vehicle-mounted camera in a raining environment can be ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of in-vehicle imaging method calibration, and particularly to a dynamic calibration method and a calibration device for an in-vehicle camera. Background Art

[0002] With the continuous development of autonomous driving technology, as one of the important sensors, the camera needs to be calibrated in real time under actual road conditions to cope with different driving scenarios and environmental changes. Dynamic calibration can correct the geometric distortion, offset, and perspective change of the camera, ensure that it obtains accurate image data, and thus improve the accuracy of target detection, obstacle recognition, and path planning.

[0003] The existing dynamic calibration methods for in-vehicle cameras mainly achieve dynamic adjustment of the internal and external parameters of the camera by analyzing the differences between the camera images and sensor data in real time, combining feature point matching, motion estimation, and optimization algorithms, so as to eliminate the deviations caused by vehicle vibration, temperature change, or mechanical displacement.

[0004] For example, the calibration device and calibration method for an in-vehicle camera disclosed in the invention patent announcement with the publication number of CN108989788B include: an automobile vehicle (10) stopped on the ground (110) within a specified stop range (111), and a display (120) disposed outside the stop range (111) in the ground (110) and displaying a calibration display pattern (130), wherein the display (120) changes the display pattern (130) according to the specifications of the automobile vehicle (10) and the in-vehicle camera (20).

[0005] For example, the calibration device for an in-vehicle camera disclosed in the invention patent announcement with the publication number of CN117459714B includes: a base, the bottom of the base is threadedly connected with a plurality of groups of legs. By placing the calibration device in front of the car and pulling out the moving box, rotating the handwheel, the rotating shaft drives the driving gear to rotate, and the driven gear drives the positioning bracket to rotate to a direction perpendicular to the ground. By pulling the pull ring, the pull rod pulls two linkage rods, so that two clamping blocks respectively enter the inside of a group of first limiting grooves. Under the tension of two groups of first torsion springs, two groups of benchmarks pop out and rotate to a position parallel to the ground. Move the base to make the positioning bracket fit the front bumper of the car, and rotate to adjust the height of the legs, so that the two groups of benchmarks are parallel to the car, so that the device is in a relatively parallel position with the vehicle when calibrating the camera, improving the calibration accuracy of the camera.

[0006] However, in the process of implementing the technical solutions of the present invention in the embodiments of the present application, it is found that the above technologies have at least the following technical problems:

[0007] In the prior art, when the vehicle camera cannot be automatically calibrated after an accident, lidar is used to assist in calibration to achieve camera position calibration. However, in rainy weather, since raindrops will reflect the laser beam emitted by the lidar, a large amount of noise is included in the signal received by the radar receiver, interfering with the detection and processing of the true echo signal by the lidar, resulting in a problem of low accuracy in camera position calibration. Summary of the Invention

[0008] The present invention provides a dynamic calibration method and a calibration device for a vehicle-mounted camera, which solve the problem in the prior art that when the vehicle camera cannot be automatically calibrated after an accident, lidar is used to assist in calibration to achieve camera position calibration. However, in rainy weather, since raindrops will reflect the laser beam emitted by the lidar, a large amount of noise is included in the signal received by the radar receiver, interfering with the detection and processing of the true echo signal by the lidar, resulting in a problem of low accuracy in camera position calibration, and improve the accuracy of automatic calibration of the camera assisted by lidar in rainy weather.

[0009] The present invention provides a dynamic calibration method for a vehicle-mounted camera, including the following steps: monitoring radar performance data in real time, and emitting a laser signal after performing radar adaptive adjustment according to the rainfall intensity in each monitoring time period; monitoring signal state data in real time, and judging whether to perform signal state optimization adjustment according to the rainfall intensity in each monitoring time period to obtain an echo signal correction value; processing each received echo signal according to the laser signal and the echo signal correction value to obtain a standard echo signal; performing quality optimization processing on the synchronously acquired camera image according to the image quality parameters, and calibrating the camera position according to the standard echo signal and the camera image after quality optimization processing.

[0010] Further, the step of monitoring radar performance data in real time and emitting a laser signal after performing radar adaptive adjustment according to the rainfall intensity in each monitoring time period includes: comprehensively analyzing the radar performance data in each monitoring time period to obtain the radar performance evaluation value in each monitoring time period; obtaining a first performance evaluation threshold and a second performance evaluation threshold from the camera calibration database; comparing the radar performance evaluation values in each monitoring time period with the first performance evaluation threshold and the second performance evaluation threshold respectively: if the radar performance evaluation value in the monitoring time period is less than the first performance evaluation threshold, turn on hardware acceleration and enter the first performance optimization mode; if the radar performance evaluation value in the monitoring time period is greater than or equal to the first performance evaluation threshold and less than the second performance evaluation threshold, enter the second performance optimization mode; if the radar performance evaluation value in the monitoring time period is greater than or equal to the second performance evaluation threshold, emit a laser signal.

[0011] Furthermore, the radar performance data includes the actual detection range, ranging accuracy, and point cloud density. The step of comprehensively analyzing the radar performance data of each monitoring time period to obtain the radar performance evaluation value of each monitoring time period includes: obtaining the critical detection range, critical ranging accuracy, critical point cloud density, and critical rainfall intensity from the camera calibration database; performing a ratio approximation operation on the actual detection range, ranging accuracy, and point cloud density of each monitoring time period with the critical detection range, critical ranging accuracy, and critical point cloud density respectively, and then performing a performance impact ratio operation, and coupling the operation results to obtain a radar performance impact parameter; performing an inverse ratio operation on the ratio approximation operation result of the average rainfall intensity of each monitoring time period and the critical rainfall intensity, and then performing a performance impact ratio operation to obtain a rainfall performance impact parameter; coupling the radar performance impact parameter and the rainfall performance impact parameter to obtain the radar performance evaluation value of each monitoring time period; and judging whether to perform radar adaptive adjustment according to the radar performance evaluation value of each monitoring time period.

[0012] Furthermore, the signal state data includes the signal-to-noise ratio, signal attenuation rate, and number of echo signals. The step of real-time monitoring the signal state data and judging whether to perform signal state optimization adjustment according to the rainfall intensity of each monitoring time period to obtain an echo signal correction value includes: obtaining the critical signal-to-noise ratio, critical signal attenuation rate, reference number of echo signals, and allowable deviation number of echo signals from the camera calibration database; performing a ratio approximation operation on the signal-to-noise ratio of each monitoring time period and the critical signal-to-noise ratio, and then performing a state impact ratio operation to obtain a signal-to-noise ratio state impact parameter; performing a ratio approximation operation on the signal attenuation rate and average rainfall intensity of each monitoring time period with the critical signal attenuation rate and critical rainfall intensity respectively, and then performing a state impact ratio operation to obtain an environmental performance impact parameter; performing a deviation compliance operation on the number of echo signals and the reference number of echo signals of each monitoring time period with the allowable deviation number of echo signals, and then performing a state impact ratio operation to obtain an echo signal number impact parameter; coupling the environmental performance impact parameter and the echo signal number impact parameter, performing an inverse ratio operation, and coupling the operation result with the signal-to-noise ratio state impact parameter to obtain the signal state evaluation value of each monitoring time period; judging whether to perform signal state optimization adjustment according to the signal state evaluation value of each monitoring time period, and obtaining an echo signal correction value according to the adjusted signal state.

[0013] Further, the steps of judging whether to perform signal state optimization adjustment according to the signal state evaluation values of each monitoring time period and obtaining the echo signal correction value according to the adjusted signal state include: obtaining a first signal state evaluation threshold and a second signal state evaluation threshold from the camera calibration database; comparing the signal state evaluation values of each monitoring time period with the first signal state evaluation threshold and the second signal state evaluation threshold respectively: if the signal state evaluation value of the monitoring time period is less than the first signal state evaluation threshold, the first resolution mode is enabled and hardware-level protection is triggered; if the signal state evaluation value of the monitoring time period is greater than or equal to the first signal state evaluation threshold and less than the second signal state evaluation threshold, adaptive filtering is started and the second resolution mode is enabled; if the signal state evaluation value of the monitoring time period is greater than or equal to the second signal state evaluation threshold, the third resolution mode is maintained; and the echo signal correction values of each monitoring time period are obtained by matching according to the adjusted signal state evaluation values of each monitoring time period.

[0014] Further, the steps of processing each received echo signal according to the laser signal and the echo signal correction value to obtain a standard echo signal include: comparing the signal frequency of each echo signal with the signal frequency of the laser signal: if the signal frequency of the echo signal is the same as the signal frequency of the laser signal, no additional processing is performed; if the signal frequency of the echo signal is different from the signal frequency of the laser signal, the echo signal is filtered to obtain each screened echo signal; the signal strength of each screened echo signal is processed by using the echo signal correction values of each monitoring time period to obtain each corrected echo signal, and it is judged whether each corrected echo signal is accurate according to the signal strength of the laser signal to obtain a standard echo signal.

[0015] Further, the steps of judging whether each corrected echo signal is accurate according to the signal strength of the laser signal to obtain a standard echo signal include: matching the signal strength of the laser signal with the echo signal strength range corresponding to each preset transmitted signal strength in the camera calibration database to obtain the echo signal strength range corresponding to the signal strength of the laser signal; judging whether the signal strength of each corrected echo signal is within the echo signal strength range, if so, marking the echo signal as a standard echo signal, otherwise adjusting the transmission frequency and re-collecting the echo signal.

[0016] Further, the picture quality parameters include image sharpness, image contrast, and noise density. The step of performing quality optimization processing on the simultaneously acquired camera pictures according to the picture quality parameters includes: obtaining the critical image sharpness, critical image contrast, critical noise density, the first picture quality threshold, and the second picture quality threshold from the camera calibration database; performing a proportion approximation operation on the image sharpness, image contrast, and critical noise density at each monitoring time point with the critical image sharpness, critical image contrast, and noise density respectively, and then performing a quality impact proportion operation on the results of the proportion approximation operation and then performing a coupling process to obtain the picture quality index at each monitoring time point; comparing the picture quality index at each monitoring time point with the first picture quality threshold and the second picture quality threshold respectively: if the picture quality index at the monitoring time point is less than the first picture quality threshold, then enter the first noise reduction mode; if the picture quality index at the monitoring time point is greater than or equal to the first picture quality threshold and less than the second picture quality threshold, then enter the second noise reduction mode; if the picture quality index at the monitoring time point is greater than or equal to the second picture quality threshold, then no additional processing is performed to obtain the processed camera picture.

[0017] Further, the step of calibrating the camera position according to the standard echo signal and the processed camera picture includes: extracting the monitoring time point and spatial coordinates of the specified feature points from the standard echo signal, and obtaining the monitoring time point and pixel coordinates of the corresponding feature points in the processed camera picture; obtaining the time offset according to the difference between the monitoring time point of the standard echo signal and the monitoring time point of the processed camera picture and the time delay data deviation processing, and the time delay data includes the signal processing time and the picture processing delay; performing time calibration on the camera according to the time offset, and at the same time converting the pixel coordinates into spatial coordinates, marking the difference between the spatial coordinates of the echo signal feature points and the spatial coordinates of the feature points in the camera picture as the spatial coordinate difference, and performing position calibration on the camera according to the spatial coordinate difference.

[0018] An embodiment of the present application provides a dynamic calibration device for a vehicle-mounted camera, including: a processor, and a memory for storing instructions executable by the processor; when the processor is configured to execute the instructions, the electronic device implements a dynamic calibration method for a vehicle-mounted camera.

[0019] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

[0020] 1. The present invention provides a dynamic calibration method and a calibration device for a vehicle-mounted camera. By adjusting the radar adaptive regulation and signal optimization according to the rainfall intensity, the accuracy and quality of the echo signal are improved. Then, by processing the frequency and correction value of the echo signal, the signal intensity of the laser signal is increased. Combining with the camera image quality parameters, the position of the camera is calibrated in real time, thereby improving the reliability of the automatic calibration method and further realizing the precise positioning and stable operation of the vehicle-mounted method in rainy environments.

[0021] 2. The present invention realizes the adaptive regulation of the radar performance by monitoring the radar performance data in real time and comprehensively analyzing the radar performance evaluation value and the rainfall intensity in each monitoring time period. Furthermore, it adapts to different radar performance requirements under different rainfall intensity conditions, ensures the best working state of the radar method under various environmental conditions, and improves the transmission efficiency and accuracy of the radar signal.

[0022] 3. The present invention monitors the signal state data in real time, optimizes the signal state according to the rainfall intensity, and further dynamically adjusts the signal state mode according to the signal state evaluation value in each monitoring time period, thereby realizing the optimization of the signal quality under different environmental conditions. Furthermore, it realizes the adaptive optimization of the signal state in complex environments and improves the stability and performance of the method.

[0023] 4. The present invention dynamically adjusts the noise reduction mode according to the camera image quality parameters to ensure that the camera can always maintain high-precision positioning under different image quality conditions. Then, it calibrates according to the time offset and spatial coordinate difference between the processed image and the standard echo signal, thereby realizing the precise calibration of the camera position. Brief Description of the Drawings

[0024] Figure 1 It is a flowchart of a dynamic calibration method for a vehicle-mounted camera provided by an embodiment of the present application.

[0025] Figure 2 It is a graph of the change of the image quality index provided by an embodiment of the present application. Detailed Embodiments

[0026] Embodiments of the present application provide a dynamic calibration method and a calibration device for an in-vehicle camera, which solve the problem in the prior art that when the vehicle camera cannot be automatically calibrated after an accident, lidar is used to assist in calibration to achieve camera position calibration. However, in rainy weather, raindrops will reflect the laser beam emitted by the radar, resulting in a large amount of noise in the signal received by the radar receiver, interfering with the detection and processing of the true echo signal by the lidar, and there is a problem of low accuracy in camera position calibration. By monitoring the radar performance data in real time and emitting laser signals after adaptive adjustment of the radar according to the rainfall intensity in each monitoring time period; monitoring the signal status data in real time and judging whether to perform signal status optimization adjustment according to the rainfall intensity in each monitoring time period to obtain an echo signal correction value; processing each received echo signal according to the laser signal and the echo signal correction value to obtain a standard echo signal; optimizing the quality of the camera image obtained synchronously according to the picture quality parameters, and calibrating the camera position according to the standard echo signal and the camera image after quality optimization processing, the accuracy of automatically calibrating the camera with the assistance of lidar in rainy weather is improved.

[0027] The technical solution in the embodiments of the present application is to solve the problem that when the vehicle camera cannot be automatically calibrated after an accident, lidar is used to assist in calibration to achieve camera position calibration. However, in rainy weather, raindrops will reflect the laser beam emitted by the radar, resulting in a large amount of noise in the signal received by the radar receiver, interfering with the detection and processing of the true echo signal by the lidar, and there is a problem of low accuracy in camera position calibration. The general idea is as follows:

[0028] By monitoring various performance indicators of the radar in real time, evaluating its working state, automatically adjusting the radar's emission parameters according to the rainfall intensity, and then judging whether it is necessary to optimize the signal state according to the real-time rainfall intensity, the influence of the rainfall environment factors on the signal is eliminated by correcting the collected echo signal to ensure the accuracy of the echo signal quality. Finally, the camera image obtained synchronously will be optimized according to the preset quality parameters, and the optimized image is compared with the standard echo signal to further calibrate the camera position, achieving the improvement of the stability and accuracy of the radar and lidar methods.

[0029] To better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings of the specification and specific embodiments.

[0030] Such as Figure 1As shown in the figure, it is a flowchart of a dynamic calibration method for an in-vehicle camera provided by an embodiment of the present application. The method includes the following steps: continuously monitor radar performance data, and emit laser signals after performing radar adaptive adjustment according to the rainfall intensity in each monitoring time period; continuously monitor signal status data, and determine whether to perform signal status optimization adjustment according to the rainfall intensity in each monitoring time period to obtain an echo signal correction value; process each received echo signal according to the laser signal and the echo signal correction value to obtain a standard echo signal; perform quality optimization processing on the synchronously acquired camera image according to the image quality parameters, and calibrate the camera position according to the standard echo signal and the camera image after quality optimization processing.

[0031] In this embodiment, since rainfall can cause attenuation and scattering of laser signals, reduce the quality of echo signals, and even generate noise signals, through radar adaptive adjustment, this method can dynamically adjust the emission parameters of laser signals according to the rainfall intensity to ensure the penetration and stability of signals in the rain; optimize the signal status according to the rainfall intensity, which can correct signal attenuation or noise caused by rainfall to obtain a more accurate echo signal correction value. The corrected echo signal can more truly reflect the distance and characteristics of the target object, improving the measurement accuracy of the radar method; through quality optimization processing, it can be ensured that even under low visibility conditions, the camera image is still clear, providing reliable visual information; by processing the echo signal to ensure a standardized echo signal is obtained, and calibrating the camera position according to the standard echo signal and the camera image after quality optimization processing, it can ensure the collaborative work of the camera and sensors such as radar, providing accurate positioning and monitoring capabilities.

[0032] In addition, the camera calibration database is used to store relevant data of a dynamic calibration method for an in-vehicle camera, including the first performance evaluation threshold, the second performance evaluation threshold, the radar performance evaluation influence factor corresponding to the actual measurement range, the critical measurement range, the critical ranging accuracy, the critical point cloud density, and the critical rainfall intensity, etc. The data in the camera calibration database can be obtained by cooperating with computer vision companies or autonomous driving companies such as Tesla, or directly obtained from public datasets such as visual benchmark test suites.

[0033] Further, the steps of real-time monitoring of radar performance data and transmitting laser signals after radar adaptive adjustment according to the rainfall intensity in each monitoring time period include: comprehensively analyzing the radar performance data in each monitoring time period to obtain the radar performance evaluation value in each monitoring time period; obtaining the first performance evaluation threshold and the second performance evaluation threshold from the camera calibration database; comparing the radar performance evaluation values in each monitoring time period with the first performance evaluation threshold and the second performance evaluation threshold respectively: if the radar performance evaluation value in the monitoring time period is less than the first performance evaluation threshold, hardware acceleration is enabled and the first performance optimization mode is entered; if the radar performance evaluation value in the monitoring time period is greater than or equal to the first performance evaluation threshold and less than the second performance evaluation threshold, the second performance optimization mode is entered; if the radar performance evaluation value in the monitoring time period is greater than or equal to the second performance evaluation threshold, laser signals are transmitted.

[0034] In this embodiment, the system will compare the radar performance evaluation value of each monitoring time period with the threshold, and select the corresponding optimization mode for the radar of each monitoring time period according to the threshold comparison result. Hardware acceleration can use GPU acceleration (graphics processing unit). In the first performance optimization mode, the transmission power, gain, filter bandwidth, and pulse bandwidth are gradually increased by 1% per single adjustment. In this embodiment, the maximum overall adjustment amplitude that can be set is 20%, which can avoid the mutation of radar performance data that may be caused by an adjustment amplitude exceeding 20%, resulting in instability or signal distortion; in the second performance optimization mode, the transmission power, gain, filter bandwidth, and pulse bandwidth are gradually increased by 1% per single adjustment. Since the method resources are limited, an adjustment amplitude of 10% can achieve a good balance between performance and resource consumption in the second performance optimization mode, so the maximum overall adjustment amplitude is 10%. By performing radar adaptive adjustment according to radar performance data and rainfall intensity, the system can flexibly respond according to the actual radar performance and environmental conditions, ensure that the radar always maintains the best working state, and provide accurate data and feedback in different situations, effectively improving the working efficiency and accuracy of the radar while reducing resource waste.

[0035] Further, the radar performance data includes the actual detection range, ranging accuracy, and point cloud density. The steps of comprehensively analyzing the radar performance data of each monitoring time period to obtain the radar performance evaluation value of each monitoring time period include: obtaining the critical detection range, critical ranging accuracy, critical point cloud density, and critical rainfall intensity from the camera calibration database; performing a ratio approximation operation on the actual detection range, ranging accuracy, and point cloud density of each monitoring time period with the critical detection range, critical ranging accuracy, and critical point cloud density respectively, and then performing a performance impact ratio operation, and coupling the operation results to obtain a radar performance impact parameter; performing an inverse ratio operation on the ratio approximation operation result of the average rainfall intensity of each monitoring time period and the critical rainfall intensity, and then performing a performance impact ratio operation to obtain a rainfall performance impact parameter; coupling the radar performance impact parameter and the rainfall performance impact parameter to obtain the radar performance evaluation value of each monitoring time period; judging whether to perform radar adaptive adjustment according to the radar performance evaluation value of each monitoring time period.

[0036] Among them, the obtaining method of the radar performance evaluation value of each monitoring time period is as follows:

[0037]

[0038] In the formula, RP i represents the radar performance evaluation value of the i-th monitoring time period, α1 represents the radar performance evaluation impact factor corresponding to the actual detection range, α2 represents the radar performance evaluation impact factor corresponding to the ranging accuracy, α3 represents the radar performance evaluation impact factor corresponding to the point cloud density, α4 represents the radar performance evaluation impact factor corresponding to the rainfall intensity, RM 1i represents the actual detection range of the i-th monitoring time period, RM0 represents the critical detection range, DA 1i represents the ranging accuracy of the i-th monitoring time period, DA0 represents the critical ranging accuracy, PC 1i represents the point cloud density of the i-th monitoring time period, PC0 represents the critical point cloud density, RI1 i represents the rainfall intensity of the i-th monitoring time period, which can be obtained by directly measuring the rainfall of the monitoring time period with a rain gauge, RI0 represents the critical rainfall intensity, where i is the number of each monitoring time period, i = 1, 2, 3,..., N, and N is the total number of monitoring time periods.

[0039] α1, α2, α3, and α4 are respectively the radar performance evaluation impact factors corresponding to the preset actual range, ranging accuracy, point cloud density, and rainfall intensity in the camera calibration database. These impact factors are numerical indicators that measure the influence of the above parameters on the radar performance evaluation value. Specifically, there is a mapping relationship table for each of the actual range, ranging accuracy, point cloud density, and rainfall intensity. The table records each possible parameter value and its corresponding radar performance evaluation impact factor. These mapping relationships can be one-to-one or many-to-one. For example, in practical applications, when it is necessary to evaluate the radar performance evaluation value for a certain monitoring time period, the measured actual range, ranging accuracy, point cloud density, and rainfall intensity can be respectively input into their corresponding mapping relationship tables, and the radar performance evaluation impact factors corresponding to these values can be quickly found. The value range of the impact factor is between 0 and 1.

[0040] In this embodiment, the actual range represents the maximum distance that the lidar can detect the echo signal when the transmit power remains unchanged within a time period, and can be calculated through the lidar equation. The lidar equation is: In the formula, P t is the laser peak power, η t is the atmospheric transmission efficiency, η r is the system efficiency (optical loss + detector quantum efficiency), ρ is the target reflectivity, A r is the receiving aperture area, P min is the minimum detectable power of the receiver, and α is the atmospheric attenuation coefficient. The ranging accuracy represents conducting multiple ranging experiments on a target object, statistically calculating the mean of the ranging results, respectively subtracting and adding the mean of the ranging results and the preset allowable deviation value to obtain the standard distance range, and statistically calculating the number of times the ranging results are within the standard distance range. The ratio of the number of times within the standard distance range to the total number of ranging experiments is the ranging accuracy; the point cloud density refers to the number of point cloud data points collected by the lidar per unit volume, which is obtained by measuring the number of laser points collected in a unit volume (such as 1 cubic decimeter). These three are interrelated. The farther the detection distance of the radar, the more obvious the attenuation of the laser signal during propagation, resulting in a decrease in the echo signal intensity, and thus the ranging accuracy may be lower; if the ranging accuracy is lower, the radar may not be able to accurately distinguish adjacent targets, resulting in a sparse point cloud density because the radar may have difficulty accurately distinguishing adjacent targets, leading to the sparsity of the point cloud data collected by the radar. The radar performance evaluation value obtained through comprehensive analysis reflects the comprehensive performance of the radar at different time periods, thus providing a basis for subsequent adaptive adjustment.

[0041] Further, the signal state data includes signal-to-noise ratio, signal attenuation rate, and the number of echo signals. The steps of monitoring the signal state data in real time and determining whether to perform signal state optimization adjustment based on the rainfall intensity in each monitoring time period to obtain the echo signal correction value include: obtaining the critical signal-to-noise ratio, critical signal attenuation rate, reference number of echo signals, and allowable deviation number of echo signals from the camera calibration database; performing a ratio approximation operation on the signal-to-noise ratio in each monitoring time period and the critical signal-to-noise ratio, and then performing a state influence ratio operation to obtain the signal-to-noise ratio state influence parameter; performing a ratio approximation operation on the signal attenuation rate and average rainfall intensity in each monitoring time period and the critical signal attenuation rate and critical rainfall intensity respectively, and then performing a state influence ratio operation to obtain the environmental performance influence parameter; performing a deviation compliance operation on the number of echo signals and the reference number of echo signals in each monitoring time period and the allowable deviation number of echo signals, and then performing a state influence ratio operation to obtain the number of echo signals influence parameter; performing a coupling process on the environmental performance influence parameter and the number of echo signals influence parameter, then performing an inverse ratio operation, and performing a coupling operation with the signal-to-noise ratio state influence parameter to obtain the signal state evaluation value for each monitoring time period; determining whether to perform signal state optimization adjustment based on the signal state evaluation value for each monitoring time period, and obtaining the echo signal correction value based on the adjusted signal state.

[0042] Among them, the method for obtaining the signal state evaluation value for each monitoring time period is as follows:

[0043]

[0044] In the formula, SS i represents the signal state evaluation value for the i-th monitoring time period, α5 represents the signal state evaluation influence factor corresponding to the signal-to-noise ratio, α6 represents the signal state evaluation influence factor corresponding to the signal attenuation rate, α7 represents the signal state evaluation influence factor corresponding to the number of echo signals, α8 represents the signal state evaluation influence factor corresponding to the rainfall intensity, SN 1i represents the average signal-to-noise ratio for the i-th monitoring time period, SN0 represents the critical signal-to-noise ratio, SA 1i represents the average signal attenuation rate for the i-th monitoring time period, SA0 represents the critical signal attenuation rate, NE 1i represents the average number of echo signals for the i-th monitoring time period, NE0 represents the reference number of echo signals, NE2 represents the allowable deviation number of echo signals, RI 1i represents the rainfall intensity for the i-th monitoring time period, and RI0 represents the critical rainfall intensity.

[0045] α5, α6, α7, and α8 are respectively the signal state evaluation impact factors corresponding to the signal-to-noise ratio, signal attenuation rate, number of echo signals, and rainfall intensity preset in the camera calibration database. These impact factors are numerical indicators for measuring the influence of the above parameters on the signal state evaluation value. Specifically, there is a mapping relationship table for each of the signal-to-noise ratio, signal attenuation rate, number of echo signals, and rainfall intensity. The table records each possible parameter value and its corresponding signal state evaluation impact factor, and these mapping relationships can be one-to-one or many-to-one. For example, in practical applications, when it is necessary to evaluate the signal state evaluation value of a certain monitoring time period, the measured signal-to-noise ratio, signal attenuation rate, number of echo signals, and rainfall intensity can be respectively input into their corresponding mapping relationship tables, and the signal state evaluation impact factors corresponding to these values can be quickly found. The value range of the impact factor is between 0 and 1.

[0046] In this embodiment, the signal-to-noise ratio refers to the ratio of the signal intensity to the noise intensity. The signal and noise powers can be measured using a signal analysis tool such as a spectrum analyzer, and the ratio of the signal power to the noise power is the signal-to-noise ratio. The higher the signal-to-noise ratio, the larger the signal state evaluation value. The signal attenuation rate is the degree of signal intensity weakening caused by factors such as distance, medium, and environment during signal propagation, and can be obtained by measuring the power difference between the source signal and the received signal using a signal analysis tool such as a spectrum analyzer. The larger the signal attenuation rate, the smaller the signal state evaluation value. The number of echo signals refers to the number of signals reflected back after the transmission signal is sent, and can be directly counted using the acquisition method of lidar. The closer the number of echo signals is to the reference number of echo signals, the larger the signal state evaluation value. The three are interrelated. The smaller the signal-to-noise ratio, the easier it is to be interfered by noise, resulting in a larger signal attenuation rate. The larger the number of echo signals, usually it means that the signal has experienced multiple reflection paths during target reflection, which may lead to a lower signal intensity at the receiving end, resulting in a smaller signal-to-noise ratio. By comprehensively analyzing the obtained signal state evaluation value, the current signal state can be accurately evaluated, and at the same time, the signal can be optimized according to the current signal state to ensure the stability and accuracy of signal transmission.

[0047] Further, the steps of judging whether to perform signal state optimization adjustment according to the signal state evaluation values of each monitoring time period and obtaining the echo signal correction value according to the adjusted signal state include: obtaining a first signal state evaluation threshold and a second signal state evaluation threshold from the camera calibration database; comparing the signal state evaluation values of each monitoring time period with the first signal state evaluation threshold and the second signal state evaluation threshold respectively: if the signal state evaluation value of the monitoring time period is less than the first signal state evaluation threshold, the first resolution mode is enabled and hardware-level protection is triggered; if the signal state evaluation value of the monitoring time period is greater than or equal to the first signal state evaluation threshold and less than the second signal state evaluation threshold, adaptive filtering is started and the second resolution mode is enabled; if the signal state evaluation value of the monitoring time period is greater than or equal to the second signal state evaluation threshold, the third resolution mode is maintained; and the echo signal correction values of each monitoring time period are obtained by matching according to the adjusted signal state evaluation values of each monitoring time period.

[0048] In this embodiment, the system compares the signal status evaluation value of each monitoring time period with a threshold, and selects a corresponding adjustment mode for the radar in each monitoring time period according to the threshold comparison result, so as to adjust the radar signal. Among them, due to the poor signal status in the first resolution mode, setting the number of vertical lines to 32 lines and the horizontal scan density to 256 lines / frame can significantly reduce the data volume and the pressure of storage and transmission. At the same time, the hardware-level protection will turn off the laser to avoid overload and send an alarm to the monitoring method; in the second resolution mode, the signal status is in the medium range. Setting the number of vertical lines to 96 lines and the horizontal scan density to 1024 lines / frame can provide finer data than the first resolution mode, support more accurate target recognition and scene reconstruction, and at the same time achieve a balance between performance and resource consumption; the signal status in the third resolution mode is good. Setting the number of vertical lines to 128 lines and the horizontal scan density to 1024 lines / frame can provide the highest quality data and at the same time provide the best performance within the scope of hardware capabilities. The echo signal correction value is a correction value obtained by adjusting the signal status evaluation value, which is used to reflect the change in the echo signal intensity after the signal optimization adjustment and can be directly obtained from the camera calibration database. For example, in the camera calibration database, a mapping relationship table is formed by corresponding the signal status evaluation value with the echo signal correction value one by one. The table records each signal status evaluation value and its corresponding echo signal correction value. These relationships can be one-to-one or many-to-one. When obtaining the echo signal correction value, only the signal status evaluation value needs to be input into the mapping relationship table, and the camera calibration database can quickly locate and return the echo signal correction value corresponding to the signal status evaluation value. By dynamically adjusting the working mode of the method according to the signal status, the present invention can effectively reduce unnecessary resource consumption, improve the working efficiency of the method. At the same time, the echo signal correction value helps to enhance the detection accuracy of the target and can better capture the target information in the echo signal.

[0049] Further, the steps of processing each received echo signal according to the laser signal and the echo signal correction value to obtain a standard echo signal include: comparing the signal frequency of each echo signal with the signal frequency of the laser signal: if the signal frequency of the echo signal is the same as the signal frequency of the laser signal, no additional processing is performed; if the signal frequency of the echo signal is different from the signal frequency of the laser signal, the echo signal is filtered to obtain each filtered echo signal; using the echo signal correction value of each monitoring time period to process the signal intensity of each filtered echo signal, adding the echo signal correction value of each monitoring time period to the signal intensity of each filtered echo signal to obtain each corrected echo signal, and judging whether each corrected echo signal is accurate according to the signal intensity of the laser signal to obtain a standard echo signal.

[0050] In this embodiment, frequency comparison helps to remove irrelevant or interfering signals, ensuring that only the echo signals matching the laser signal frequency are retained, thereby improving the accuracy and reliability of signal processing. Through the correction of signal intensity, the possible attenuation or distortion of the echo signal during transmission can be compensated, ensuring that the signal intensity meets the expectations, and further improving the signal quality and accuracy. Through accuracy judgment, it can be ensured that the finally obtained echo signal is valid and accurate, avoiding measurement deviations caused by inaccurate or incorrect signals, and further improving the precision and reliability of the method. The present invention can effectively filter out unnecessary interference and noise, obtain cleaner and clearer echo signals, help improve the signal quality of the method, and ensure the precision and reliability of subsequent processing; the corrected echo signal matches the laser signal in both intensity and frequency, which can reduce errors caused by inaccurate signals, thereby obtaining more accurate target detection and measurement results.

[0051] Further, the steps of judging whether each corrected echo signal is accurate according to the signal intensity of the laser signal and obtaining the standard echo signal include: matching the signal intensity of the laser signal with the echo signal intensity ranges corresponding to the preset transmitted signal intensities in the camera calibration database to obtain the echo signal intensity range corresponding to the signal intensity of the laser signal; judging whether the signal intensity of each corrected echo signal is within the echo signal intensity range, and if so, marking the echo signal as the standard echo signal, otherwise adjusting the transmission frequency and re-collecting the echo signal.

[0052] In this embodiment, the echo signal intensity range is a reasonable interval of the echo signal intensity corresponding to a specific transmitted signal intensity, which can be directly obtained from the camera calibration database. For example, in the camera calibration database, a mapping relationship table is formed by corresponding the signal intensity of the laser signal with the echo signal intensity range one by one. The table records the echo signal intensity range corresponding to each transmitted signal intensity, and these relationships can be one-to-one or many-to-one. When obtaining the echo signal intensity range of the laser signal, just input the signal intensity into the mapping relationship table, and the camera calibration database can quickly locate and return the echo signal intensity range corresponding to the transmitted signal intensity. By matching the signal intensity of the laser signal with the preset echo signal intensity range, the echo signals meeting the standards can be effectively screened out, thereby improving the measurement accuracy. Judging whether the corrected echo signal is within the preset range can exclude abnormal signals caused by environmental interference or equipment errors, reducing measurement errors. If the echo signal is not within the preset range, the method will automatically adjust the transmission frequency and re-collect the signal, reducing the need for manual intervention and improving the automation degree and efficiency of the method.

[0053] Further, the picture quality parameters include image sharpness, image contrast, and noise density. The steps for optimizing the quality of the camera pictures obtained synchronously according to the picture quality parameters include: obtaining the critical image sharpness, critical image contrast, critical noise density, first picture quality threshold, and second picture quality threshold from the camera calibration database; performing a ratio approximation operation on the image sharpness, image contrast, and critical noise density at each monitoring time point with the critical image sharpness, critical image contrast, and noise density respectively, and then performing a quality impact ratio operation on the results of the ratio approximation operation and then performing a coupling process to obtain the picture quality index at each monitoring time point; comparing the picture quality index at each monitoring time point with the first picture quality threshold and the second picture quality threshold respectively: if the picture quality index at the monitoring time point is less than the first picture quality threshold, enter the first noise reduction mode; if the picture quality index at the monitoring time point is greater than or equal to the first picture quality threshold and less than the second picture quality threshold, enter the second noise reduction mode; if the picture quality index at the monitoring time point is greater than or equal to the second picture quality threshold, no additional processing is performed, and the processed camera picture is obtained.

[0054] Among them, the method for obtaining the picture quality index at each monitoring time point is as follows:

[0055]

[0056] In the formula, IQ j represents the picture quality index at the j-th monitoring time point, β1 represents the picture quality index impact factor corresponding to image sharpness, β2 represents the picture quality index impact factor corresponding to image contrast, β3 represents the picture quality index impact factor corresponding to noise density, Cl 1j represents the image sharpness at the j-th monitoring time point, Cl0 represents the critical image sharpness, Co 1j represents the image contrast at the j-th monitoring time point, Co0 represents the critical image contrast, ND 1j represents the noise density at the j-th monitoring time point, ND0 represents the critical noise density, where j is the number of each monitoring time point, j = 1, 2, 3,..., M, and M is the total number of monitoring time points.

[0057] β1, β2, and β3 are respectively the picture quality index impact factors corresponding to the preset image sharpness, image contrast, and noise density in the camera calibration database. These impact factors are numerical indicators that measure the influence of the above-mentioned picture quality parameters on the picture quality index. Specifically, there is a mapping relationship table for each of image sharpness, image contrast, and noise density. The table records each possible picture quality parameter value and its corresponding picture quality index impact factor. These mapping relationships can be one-to-one or many-to-one. For example, in practical applications, when it is necessary to evaluate the picture quality index at a certain monitoring time point, the measured image sharpness, image contrast, and noise density can be respectively input into their corresponding mapping relationship tables, and the picture quality index impact factors corresponding to these values can be quickly found. The value range of the impact factor is between 0 and 1.

[0058] In this embodiment, sharpness represents the clarity of image details. It is obtained by calculating the Laplacian operator of the image to measure the sharpness of the edges and then calculating the variance of the Laplacian operator of the image. The method for obtaining the Laplacian operator is as follows: Among them, represents the Laplacian operator of the image, I(x, y) represents the pixel value of the image at position (x, y), represents the second-order partial derivative of the image I(x, y) with respect to the x direction, reflecting the change of pixel values in the horizontal direction, represents the second-order partial derivative of the image I(x, y) with respect to the y direction, reflecting the change of pixel values in the vertical direction; contrast reflects the difference between the brightest and darkest regions in the image and can be obtained by calculating the difference between the maximum brightness value and the minimum brightness value of the image pixel points; noise density represents the degree of noise distribution in the image and can be directly obtained by using noise monitoring algorithms such as median filtering to analyze the changes in local regions of the image. Among them, the three are interrelated. When the noise density of the image is relatively high, the noise usually blurs the details, resulting in low image sharpness; an image with higher contrast can highlight the details more and reduce the visual interference of noise, but if the noise density is too large, it will still affect the overall image effect. The picture quality index obtained through comprehensive analysis can determine whether there are problems such as blurring, excessive noise, or uneven contrast in the image, and adjust or optimize it in a timely manner. The method can evaluate the quality level of the image in real time to avoid the situation of unqualified image quality affecting subsequent processing or decision-making.

[0059] It should be added that the system will compare the picture quality index at each monitoring time point with a threshold, and select the corresponding quality optimization mode for the camera picture at each monitoring time point according to the result of the threshold comparison. When entering the first noise reduction mode, advanced noise reduction algorithms such as denoising convolutional neural networks are used to reduce noise in the camera picture. At the same time, the exposure value is reduced by 2 stops, and the contrast is increased by 30%. If the picture quality index after processing is still less than the first picture quality threshold, image restoration technology is used to fill in the missing details or textures. When entering the second noise reduction mode, noise reduction algorithms such as median filtering and Gaussian filtering are used to remove the noise introduced by water vapor or raindrops in rainy environments. At the same time, the exposure value is reduced by 1 stop, and the contrast is increased by 20%. By dynamically adapting the adjustment according to the picture quality index, the processing mode can be adjusted according to the current picture quality, which can ensure that the picture is always in the best state. Whether in an environment with poor light or in a situation with good picture quality, the method can select the most suitable processing mode according to the actual situation, ensuring the stable operation of the method. By comparing the image after noise reduction processing with the standard echo signal, the position of the camera can be calibrated more accurately.

[0060] Set the picture quality index influence factor corresponding to image sharpness to 0.3, the picture quality index influence factor corresponding to image contrast to 0.3, the picture quality index influence factor corresponding to noise density to 0.4. Set the image sharpness at the j-th monitoring time point to 1.2, the critical image sharpness to 1.5, the image contrast at the j-th monitoring time point to 0.6, the critical image contrast to 0.5, and the critical noise density to 0.5. In the case of continuous increase in noise density, calculate the picture quality index at the j-th monitoring time point. As shown in Table 1, the data table of the picture quality index of the dynamic calibration method based on in-vehicle cameras.

[0061] Table 1 Data table of the picture quality index of the dynamic calibration method based on in-vehicle cameras

[0062] Number <![CDATA[ND 1j > <![CDATA[IQ j > 1 0.3 1.267 2 0.4 1.1 3 0.5 1 4 0.6 0.933 5 0.7 0.886

[0063] As Figure 2 shown, it is the change diagram of the picture quality index of the dynamic calibration method based on in-vehicle cameras provided by the embodiment of the present application. As shown in Table 1 and Figure 2 it can be seen that when the picture quality index influence factor corresponding to image sharpness, the picture quality index influence factor corresponding to image contrast, the picture quality index influence factor corresponding to noise density, the image sharpness at the j-th monitoring time point, the critical image sharpness, the image contrast at the j-th monitoring time point, the critical image contrast, and the critical noise density remain unchanged, and the noise density continuously increases, the picture quality index at the j-th monitoring time point will continuously decrease.

[0064] Further, the steps of calibrating the camera position according to the standard echo signal and the camera image after quality optimization processing include: extracting the monitoring time points and spatial coordinates of the specified feature points from the standard echo signal, and obtaining the monitoring time points and pixel coordinates of the corresponding feature points in the processed camera image; obtaining the time offset according to the difference between the monitoring time point of the standard echo signal and the monitoring time point of the processed camera image and the time delay data deviation processing. The time delay data includes the signal processing time and the image processing delay; performing time calibration on the camera according to the time offset, and at the same time converting the pixel coordinates into spatial coordinates, marking the difference between the spatial coordinates of the echo signal feature points and the spatial coordinates of the feature points in the camera image as the spatial coordinate difference, and performing position calibration on the camera according to the spatial coordinate difference.

[0065] In this embodiment, the time offset is obtained as follows: TO = ΔT - T sp -T ip ; where TO represents the time offset, ΔT represents the difference between the monitoring time point of the standard echo signal and the monitoring time point of the processed camera image, and T sp represents the signal processing time, which can be directly obtained by using a hardware timer to record the time elapsed from the start of signal acquisition to the completion of processing and output. T ip represents the image processing delay, which can be directly obtained by using a hardware timer to record the time required from camera shooting to image display. Applying the time offset to the camera data stream to align the image timestamp with the echo signal can complete the time calibration. When performing position calibration, the conversion from pixel coordinates to spatial coordinates is usually completed through the perspective projection model of the camera. The camera projection model describes the conversion relationship from the world coordinate system to the camera coordinate system and then to the pixel coordinates. The projection model of the camera can be described by the following equation: where [u, v] are the pixel coordinates of the image, K is the camera internal parameter matrix, which contains focal length and principal point information, [R|t] is the camera external parameter matrix, which contains the rotation matrix R and the translation vector t, and [X, Y, Z] are the three-dimensional spatial coordinates of the camera in the world coordinate system. When converting pixel coordinates to spatial coordinates, the point in the camera coordinate system can be calculated from the pixel coordinates [u, v] according to the camera internal parameter matrix K: where [x c , y c , z c is the point in the camera coordinate system, and K -1 is the inverse matrix of the camera internal parameter matrix. Then, the point in the camera coordinate system is converted into the spatial coordinates [X c , Y c , Z c , and the specific method is: In the formula, f x and f y are the horizontal and vertical focal lengths of the camera respectively, and Z is the distance from the feature point to the camera. Finally, through the external parameter matrix [R|t] of the camera, the three-dimensional coordinates in the camera coordinate system are converted into the coordinates in the world coordinate system: In the formula, [X w , Y w , Z w represents the spatial coordinates of the corresponding feature point in the processed camera image, R is the rotation matrix of the camera, and t is the translation vector of the camera.

[0066] An embodiment of the present application further provides a dynamic calibration device for a vehicle-mounted camera, including: a processor and a memory for storing instructions executable by the processor; when the processor is configured to execute the instructions, the electronic device implements a dynamic calibration method for a vehicle-mounted camera.

[0067] In summary, in the embodiment of the present application, by real-time monitoring of various performance indicators of the radar, evaluating its working state, automatically adjusting the transmission parameters of the radar according to the rainfall intensity, and then judging whether it is necessary to optimize and adjust the signal state according to the real-time rainfall intensity. Once the echo signal is collected, the echo signal is corrected to eliminate the influence of environmental factors such as rainfall on the signal, ensuring the accuracy of the echo signal quality. Finally, the obtained camera image is optimized according to the preset quality parameters, and the optimized image is compared with the standard echo signal to further calibrate the position of the camera, achieving the improvement of the stability and accuracy of the radar and laser methods.

[0068] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a method, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0069] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (methods), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, and the combination of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate for implementing in the process Figure 1One or more processes and / or blocks Figure 1 Apparatus for the functions specified in one or more blocks

[0070] These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce a manufacture including an instruction apparatus that implements the functions in the process Figure 1 One or more processes and / or blocks Figure 1 The functions specified in one or more blocks

[0071] These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus, such that a series of operational steps are performed on the computer or other programmable apparatus to produce a computer-implemented process, whereby the instructions executed on the computer or other programmable apparatus provide steps for implementing the functions in the process Figure 1 One or more processes and / or blocks Figure 1 The steps of the functions specified in one or more blocks

[0072] Although the preferred embodiments of the present invention have been described, additional changes and modifications can be made by those skilled in the art once they learn of the basic inventive concept. Therefore, the appended claims are intended to be construed to include the preferred embodiments as well as all changes and modifications that fall within the scope of the present invention

[0073] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these modifications and variations

Claims

1. A dynamic calibration method for an in-vehicle camera, characterized in that Including the following steps: Real-time monitor the radar performance data, and emit a laser signal after performing radar adaptive adjustment according to the rainfall intensity of each monitoring time period; Real-time monitor the signal status data, and judge whether to perform signal status optimization adjustment according to the rainfall intensity of each monitoring time period to obtain an echo signal correction value; Process each received echo signal according to the laser signal and the echo signal correction value to obtain a standard echo signal; Perform quality optimization processing on the camera images obtained synchronously according to the image quality parameters, and calibrate the camera position according to the standard echo signal and the camera images after quality optimization processing.

2. The dynamic calibration method of an in-vehicle camera according to claim 1, characterized in that: The step of real-time monitoring the radar performance data and emitting a laser signal after performing radar adaptive adjustment according to the rainfall intensity of each monitoring time period includes: Comprehensively analyze the radar performance data of each monitoring time period to obtain the radar performance evaluation value of each monitoring time period; Obtain the first performance evaluation threshold and the second performance evaluation threshold from the camera calibration database; Compare the radar performance evaluation values of each monitoring time period with the first performance evaluation threshold and the second performance evaluation threshold respectively: If the radar performance evaluation value of the monitoring time period is less than the first performance evaluation threshold, turn on hardware acceleration and enter the first performance optimization mode; If the radar performance evaluation value of the monitoring time period is greater than or equal to the first performance evaluation threshold and less than the second performance evaluation threshold, enter the second performance optimization mode; If the radar performance evaluation value of the monitoring time period is greater than or equal to the second performance evaluation threshold, emit a laser signal.

3. The dynamic calibration method of a vehicle-mounted camera according to claim 2, characterized in that: The radar performance data includes the actual measurement range, ranging accuracy, and point cloud density; The step of comprehensively analyzing the radar performance data of each monitoring time period to obtain the radar performance evaluation value of each monitoring time period includes: Obtain the critical measurement range, critical ranging accuracy, critical point cloud density, and critical rainfall intensity from the camera calibration database; Perform a ratio approximation operation on the actual measurement range, ranging accuracy, and point cloud density of each monitoring time period with the critical measurement range, critical ranging accuracy, and critical point cloud density respectively, and then perform a performance impact ratio operation, and perform a coupling process on the operation results to obtain a radar performance impact parameter; Perform an inverse ratio operation on the ratio approximation operation result of the average rainfall intensity of each monitoring time period and the critical rainfall intensity, and then perform a performance impact ratio operation on the operation result to obtain a rainfall performance impact parameter; Perform a coupling process on the radar performance impact parameter and the rainfall performance impact parameter to obtain the radar performance evaluation value of each monitoring time period; Judge whether to perform radar adaptive adjustment according to the radar performance evaluation value of each monitoring time period.

4. The dynamic calibration method of a vehicle-mounted camera according to claim 1, wherein: The signal status data includes the signal-to-noise ratio, signal attenuation rate, and number of echo signals; The step of real-time monitoring the signal status data and judging whether to perform signal status optimization adjustment according to the rainfall intensity of each monitoring time period to obtain an echo signal correction value includes: Obtain the critical signal-to-noise ratio, critical signal attenuation rate, reference number of echo signals, and allowable deviation number of echo signals from the camera calibration database; Perform a proportion approximation operation on the signal-to-noise ratio of each monitoring time period and the critical signal-to-noise ratio, and then perform a state influence proportion operation to obtain a signal-to-noise ratio state influence parameter; Perform a proportion approximation operation on the signal attenuation rate and average rainfall intensity of each monitoring time period and the critical signal attenuation rate and critical rainfall intensity respectively, and then perform a state influence proportion operation to obtain an environmental performance influence parameter; Perform a deviation compliance operation on the number of echo signals and the reference number of echo signals of each monitoring time period and the allowable deviation number of echo signals, and then perform a state influence proportion operation to obtain an echo signal number influence parameter; Couple-process the environmental performance influence parameter and the echo signal number influence parameter, perform an inverse proportion operation, and perform a coupling operation with the signal-to-noise ratio state influence parameter to obtain a signal state evaluation value for each monitoring time period; Judge whether to perform signal state optimization adjustment according to the signal state evaluation value of each monitoring time period, and obtain an echo signal correction value according to the adjusted signal state.

5. The dynamic calibration method of an in-vehicle camera according to claim 4, characterized in that: The step of judging whether to perform signal state optimization adjustment according to the signal state evaluation value of each monitoring time period and obtaining an echo signal correction value according to the adjusted signal state includes: Obtain a first signal state evaluation threshold and a second signal state evaluation threshold from the camera calibration database; Compare the signal state evaluation value of each monitoring time period with the first signal state evaluation threshold and the second signal state evaluation threshold respectively: If the signal state evaluation value of the monitoring time period is less than the first signal state evaluation threshold, turn on the first resolution mode and trigger hardware-level protection; If the signal state evaluation value of the monitoring time period is greater than or equal to the first signal state evaluation threshold and less than the second signal state evaluation threshold, start adaptive filtering and turn on the second resolution mode; If the signal state evaluation value of the monitoring time period is greater than or equal to the second signal state evaluation threshold, maintain the third resolution mode; Match and obtain an echo signal correction value for each monitoring time period according to the signal state evaluation value after adjustment of each monitoring time period.

6. The dynamic calibration method of an in-vehicle camera according to claim 1, characterized in that: The step of processing each received echo signal according to the laser signal and the echo signal correction value to obtain a standard echo signal includes: Compare the signal frequency of each echo signal with the signal frequency of the laser signal: If the signal frequency of the echo signal is the same as the signal frequency of the laser signal, no additional processing is performed; If the signal frequency of the echo signal is different from the signal frequency of the laser signal, filter the echo signal to obtain each screened echo signal; Process the signal intensity of each screened echo signal by using the echo signal correction value of each monitoring time period to obtain each corrected echo signal, and judge whether each corrected echo signal is accurate according to the signal intensity of the laser signal to obtain a standard echo signal.

7. The dynamic calibration method of an in-vehicle camera according to claim 6, characterized in that: The step of judging whether each corrected echo signal is accurate according to the signal intensity of the laser signal to obtain a standard echo signal includes: Match the signal intensity of the laser signal with the echo signal intensity range corresponding to each preset transmitted signal intensity in the camera calibration database to obtain the echo signal intensity range corresponding to the signal intensity of the laser signal; Determine whether the signal intensity of each echo signal after correction is within the range of the echo signal intensity. If so, mark the echo signal as a standard echo signal; otherwise, adjust the transmission frequency and re-acquire the echo signal.

8. The dynamic calibration method of an in-vehicle camera according to claim 1, characterized in that: The picture quality parameters include image sharpness, image contrast, and noise density. The step of performing quality optimization processing on the synchronously acquired camera picture according to the picture quality parameters includes: Obtain the critical image sharpness, critical image contrast, critical noise density, the first picture quality threshold, and the second picture quality threshold from the camera calibration database. Perform ratio approximation operations on the image sharpness, image contrast, and critical noise density at each monitoring time point with the critical image sharpness, critical image contrast, and noise density respectively, and then perform quality impact ratio operations on the results of the ratio approximation operations and then perform coupling processing to obtain the picture quality index at each monitoring time point. Compare the picture quality index at each monitoring time point with the first picture quality threshold and the second picture quality threshold respectively: If the picture quality index at the monitoring time point is less than the first picture quality threshold, enter the first noise reduction mode. If the picture quality index at the monitoring time point is greater than or equal to the first picture quality threshold and less than the second picture quality threshold, enter the second noise reduction mode. If the picture quality index at the monitoring time point is greater than or equal to the second picture quality threshold, no additional processing is performed, and the processed camera picture is obtained.

9. The dynamic calibration method of an in-vehicle camera according to claim 1, characterized in that: The step of calibrating the camera position according to the standard echo signal and the camera picture after quality optimization processing includes: Extract the monitoring time point and spatial coordinates of the specified feature point from the standard echo signal, and obtain the monitoring time point and pixel coordinates of the corresponding feature point in the processed camera picture. Obtain the time offset according to the difference between the monitoring time point of the standard echo signal and the monitoring time point of the processed camera picture and the time delay data deviation processing. The time delay data includes signal processing time and picture processing delay. Perform time calibration on the camera according to the time offset, and at the same time convert the pixel coordinates into spatial coordinates. Mark the difference between the spatial coordinates of the echo signal feature point and the spatial coordinates of the feature point in the camera picture as the spatial coordinate difference, and perform position calibration on the camera according to the spatial coordinate difference.

10. A dynamic calibration device for a vehicle-mounted camera, characterized in that, Include: A processor and a memory for storing instructions executable by the processor. When the processor is configured to execute the instructions, the electronic device implements a dynamic calibration method for an in-vehicle camera as described in any one of claims 1-9.

Images