A synchronous calibration system for multi-angle train image acquisition

Abstract

The present invention relates to the field of train image acquisition, and specifically to a synchronous calibration system for multi-angle train image acquisition, comprising an image acquisition module, a calibration module, a first deviation comparison module, a second deviation comparison module, a verification module, a calibration module and a cloud database. By using a multi-angle image acquisition device to overlap and acquire images, and using calibration points to compare image deviations and verify, efficient and accurate calibration of the train image acquisition device is achieved, and the accuracy and reliability of image analysis are improved, thereby providing beneficial effects of technical support for train inspection, maintenance and the like.

Description

Technical Field

[0001] The invention relates to the field of train image acquisition, and in particular to a synchronous calibration system for multi-angle train image acquisition. Background Art

[0002] Train inspection plays an important role in rail train safety. However, in the prior art, train images need to be synchronously collected before train inspection. Conventional train synchronous image acquisition systems may have errors during use. Therefore, in order to solve the problems of asynchronous multi-angle train image acquisition and errors in the acquisition range, the present invention proposes a synchronous calibration system for multi-angle train image acquisition. By using overlapping acquired images and comparing image deviations with calibration points and verifying them, efficient and accurate calibration of the train image acquisition device is achieved, thereby improving the accuracy and reliability of image analysis, and providing technical support for train inspection, maintenance and other tasks. Summary of the invention

[0003] In order to solve the problems existing in the above-mentioned prior art, the first purpose of the present invention is to provide a synchronous calibration system for multi-angle train image acquisition. By using overlapping acquired images and applying calibration points to compare image deviations and verify them, efficient and accurate calibration of the train image acquisition device is achieved, thereby improving the accuracy and reliability of image analysis and providing technical support for train inspection, maintenance and other tasks.

[0004] In order to achieve the above object, the present invention provides a solution: a synchronous calibration system for multi-angle train acquisition images, comprising:

[0005] The image acquisition module includes a plurality of image acquisition devices for acquiring images to be tested; the image acquisition module is composed of a plurality of image acquisition devices, and the image acquisition devices are distributed at different positions of the train detection device for capturing all-round images of the train.

[0006] The calibration module is fixedly connected to the image acquisition module and is used to map calibration points to the area to be measured within the sampling range of the image acquisition module; the calibration module is tightly connected to the image acquisition module and is used to set calibration points within the image acquisition range, and the calibration points always maintain the same connection with the connected image acquisition device, thereby ensuring the accuracy of subsequent image deviation comparison and calibration.

[0007] The first deviation comparison module generates hue deviation information by comparing the hue deviation between the standard acquired image and the image to be tested. By comparing the hue deviation of the image to be tested, it is possible to detect whether there is hue and color temperature deviation in the train image, and perform hue correction, thereby providing a more accurate image to be tested for subsequent train anomaly detection.

[0008] The second deviation comparison module compares whether there is a sampling deviation between image acquisition devices by using a standard acquisition image to obtain sampling deviation information, and associates the image acquisition device with the sampling deviation information to obtain device deviation information.

[0009] The verification module is used to detect the accuracy of the color tone deviation information, the sampling deviation information and the device deviation information; the verification module can ensure the accuracy and credibility of the comparison result and avoid misjudgment or missed judgment.

[0010] A calibration module receives hue deviation information, sampling deviation information and device deviation information, calls hue calibration parameters corresponding to the hue deviation information, image calibration parameters corresponding to the sampling deviation information and device calibration parameters corresponding to the device deviation information in the cloud database, performs image hue calibration and image position calibration on the image to be tested, and performs hardware calibration on the image acquisition device; the calibration module can calibrate images and image acquisition devices with deviations, thereby eliminating the subsequent impact of sampling deviations on train detection.

[0011] The cloud database is used to store standard captured images, tone calibration parameters, image calibration parameters and device calibration parameters, and can quickly judge and calibrate image deviations, thereby improving image processing efficiency.

[0012] As a further improvement of the present invention, the calibration module is provided with a laser ray lamp, which is fixedly connected to the image acquisition device. The laser ray lamp generates a calibration point with a fixed position in the image to be measured. By using the laser ray lamp fixedly connected to the image acquisition device to perform calibration point mapping, the calibration point position can be always at a fixed position in the acquired image. When the position and angle of the image acquisition device are changed, the calibration module is changed together with it, thereby ensuring that the position of the calibration point in the image remains unchanged. The image acquisition devices are evenly distributed on all sides of the train, ensuring all-round image acquisition of the train from multiple angles, and being able to capture all details of the train, providing a rich data source for subsequent image analysis and processing. There are overlapping sampling areas between adjacent image acquisition devices, and the calibration points fall within the overlapping sampling areas on each side of the image acquisition device, ensuring that there is sufficient overlapping part between images taken at different angles. The overlapping part can be used for subsequent image deviation comparison and calibration and other processing processes to improve the overall consistency and accuracy of the image.

[0013] As a further improvement of the present invention, the first deviation comparison module performs denoising on the image to be tested and compares the hue components of the color space in the image, calculates the hue components of the image to be tested and the hue components of the standard acquired image, reflects the hue difference between the image to be tested and the standard image according to the hue comparison result, generates hue deviation information, and by comparing the hue deviation of the image to be tested, detects whether the train image has hue and color temperature deviations, and performs hue correction, thereby providing a more accurate image to be tested for subsequent train abnormality detection.

[0014] As a further improvement of the present invention, the second deviation comparison module selects two adjacent image acquisition devices in turn, and obtains the image to be tested and the auxiliary deviation comparison image through the two image acquisition devices respectively. The second deviation comparison module identifies the coordinates of the calibration points in the acquired image to be tested and the auxiliary deviation comparison image, and calculates the relative position spacing value of the calibration points of the image to be tested and the auxiliary deviation comparison image and the relative position spacing value difference of the calibration points in the standard acquisition image to obtain the calibration point deviation parameter K; the relative position spacing value refers to the pixel distance or actual distance between the corresponding calibration points in the two images, and the standard spacing value is obtained by precise measurement in the system calibration stage, representing the ideal spacing between the calibration points in the images acquired by the adjacent image acquisition devices without deviation. By accurately identifying the coordinates of the calibration points in the image to be tested and comparing the relative position spacing values ​​of the calibration points in the overlapping sampling area in the images acquired by the adjacent image acquisition devices, the possible deviations between the image acquisition devices can be discovered and quantified in time. This method based on calibration points ensures the accuracy and reliability of deviation detection and provides a solid foundation for subsequent calibration work.

[0015] The second deviation comparison module also extracts the overlapping sampling areas in the image to be tested and the auxiliary deviation comparison image acquired by the image acquisition device through image feature matching, and deletes the non-overlapping areas in the image to obtain the overlapping sampling images; the standard acquisition image also includes a standard overlapping image, and the standard overlapping image is an image of the overlapping area of ​​the standard acquisition image and the image acquired by the adjacent image acquisition device; the overlapping deviation parameter S is obtained by comparing the difference in the length of the four sides and the difference in the area between the overlapping sampling image and the standard overlapping image; by comparing the overlapping areas of the image to be tested and the adjacent images with the parameters of the overlapping areas in the standard acquisition image, a new reference factor is provided for the deviation comparison, thereby improving the accuracy of the deviation comparison.

[0016] The second deviation comparison module performs pattern recognition, identifies the marker pattern in the image to be tested, and identifies the matching marker pattern in the standard acquisition image. By comparing the matching degree of the marker pattern in the image to be tested with the marker pattern in the standard acquisition image, the difference between the matching degree and the complete matching degree is calculated to obtain the pattern abnormality deviation parameter M; by performing feature matching on the pattern with obvious image features in the image, accurate comparison factors can be provided for deviation comparison.

[0017] The cloud database also stores a deviation threshold. The second deviation comparison module calculates the sampling deviation value by combining the calibration point deviation parameter K, the overlap deviation parameter S and the pattern abnormality deviation parameter M, and compares the sampling deviation value with the deviation threshold to obtain the sampling deviation information.

[0018] By considering the coordinates of the calibration points, coincident sampling and characteristic pattern factors, the relative positional relationship between image acquisition devices can be more comprehensively evaluated, thereby further improving the comparison accuracy and ensuring the reliability and accuracy of the results. The introduction of the second deviation comparison module makes the calibration process more complete and systematic. It not only focuses on the deviation between adjacent image acquisition devices, but also further verifies the calibration effect through cross-side comparison, providing higher quality data support for subsequent image analysis and processing.

[0019] As a further improvement of the present invention, the calibration point deviation parameter is calculated as follows:

[0020] ;

[0021] ;

[0022] Among them, k1 is the relative position spacing value of the calibration points of the image to be tested and the auxiliary deviation comparison image, (x1, y1) is the coordinate of the calibration point in the image to be tested, (x2, y2) is the coordinate of the calibration point in the auxiliary deviation comparison image, and k0 is the relative position spacing value difference of the calibration points in the standard acquisition image. The smaller K is, the smaller the possibility of deviation.

[0023] As a further improvement of the present invention, the overlap deviation parameter is calculated as follows:

[0024] ;

[0025] Among them, l i is the length of the i-th side of the overlapped sampled image, s is the area of ​​the overlapped sampled image, l0 i is the length of the i-th side of the standard coincidence image, and s0 is the area of ​​the standard coincidence image. By using the coincidence deviation parameter, an accurate comparison factor is provided for the verification of the second deviation comparison module, thereby improving the accuracy of the deviation verification.

[0026] As a further improvement of the present invention, the sampling deviation values ​​are:

[0027] ;

[0028] Among them, Q is the sampling deviation value, α is the coefficient of the calibration point deviation parameter, β is the coefficient of the overlap deviation parameter, and γ is the coefficient of the pattern abnormality deviation parameter; Q0 is the deviation threshold, and the sampling deviation information is obtained by comparing the values ​​of Q and Q0. By setting the deviation threshold, when the sampling deviation value is greater than the deviation threshold, it is judged that the image to be tested has a sampling deviation. By using the deviation threshold, the train sampling deviation comparison accuracy in different application scenarios can be set.

[0029] As a further improvement of the present invention, the verification module verifies the accuracy of the hue deviation information of the image to be tested by comparing the hue deviation between the image to be tested and the adjacent auxiliary deviation comparison image; the image acquisition device also acquires train images from multiple angles, and the verification module detects the accuracy of the sampling deviation information and the device deviation information of the image to be tested by feature matching the image to be tested and the train image. The existence of the verification module ensures the accuracy and reliability of the entire image acquisition and calibration system. By timely verifying the accuracy of the deviation comparison, potential deviation problems can be discovered and corrected in a timely manner, avoiding these deviations from having an adverse effect on subsequent image analysis and processing. At the same time, the verification module also provides an important basis for the maintenance and optimization of the system, which helps to improve the overall performance and stability of the system.

[0030] As a further improvement of the present invention, the calibration module receives the hue deviation information and the sampling deviation information, and matches the corresponding hue calibration parameters and sampling deviation parameters in the cloud database to calibrate the hue and image position of the image to be tested. The first-level deviation value is matched to the corresponding calibration parameters in the cloud database, and the first-level calibration is performed on the acquired image to be tested, and the deviation of the calibrated image to be tested is compared and verified again. The calibration module can ensure the accuracy and reliability of the image acquisition and calibration system. By timely calibrating the hue, color temperature and sampling position of the image, the deviation in the image can be eliminated or reduced, and the quality and consistency of the image can be improved.

[0031] As a further improvement of the present invention, the image acquisition device is provided with a rotation mechanism, the calibration module receives the device deviation information, and matches the corresponding device calibration parameters in the cloud database through the device deviation information, and adjusts the angle and position of the image acquisition device through the device calibration parameters to perform hardware calibration; when the calibration module performs hardware calibration, the calibration module connected to the image acquisition device adjacent to the calibrated image acquisition device emits rays to the target correction position in the overlapping sampling area to form a target calibration point, and adjusts the rotation mechanism of the calibrated image acquisition device so that the calibration point emitted by the calibration module connected to it coincides with the target calibration point to complete the hardware calibration. By using the calibration point to calibrate the device deviation, the accuracy of the calibration can be ensured, and by calibrating the device deviation, the deviation of the image acquisition device can be calibrated in time, ensuring the accuracy of subsequent image acquisition.

[0032] The beneficial effects of the present invention are:

[0033] The synchronous calibration system for multi-angle train image acquisition proposed in the present invention verifies the offset of the angle of the image acquisition device by using a deviation comparison module and a calibration module, thereby solving the problem of poor image stitching effect and inaccurate detection area caused by asynchronous image acquisition in the pre-image acquisition step of train detection in the prior art, thereby improving the stability of overall train detection, realizing efficient and accurate calibration of the train image acquisition device, improving the accuracy and reliability of image analysis, and providing technical support for train detection, maintenance and other tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is a system block diagram of a synchronous calibration system for multi-angle train acquisition images of the present invention;

[0035] Figure 2 It is a schematic diagram of sampling and calibration of multi-angle train acquisition images according to an embodiment of the present invention. DETAILED DESCRIPTION

[0036] The specific embodiments of the present invention will be described in detail below. It should be noted that the embodiments described here are only part of the embodiments of the present invention, not all of the embodiments. Based on the following embodiments, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0037] The embodiment of the present invention provides a synchronous calibration system for multi-angle train acquisition images, comprising:

[0038] The image acquisition module includes a plurality of image acquisition devices for acquiring images to be tested; the image acquisition module is composed of a plurality of image acquisition devices, and the image acquisition devices are distributed at different positions of the train detection device for capturing all-round images of the train.

[0039] The calibration module is fixedly connected to the image acquisition module and is used to map calibration points to the area to be measured within the sampling range of the image acquisition module; the calibration module is tightly connected to the image acquisition module and is used to set calibration points within the image acquisition range, and the calibration points always maintain the same connection with the connected image acquisition device, thereby ensuring the accuracy of subsequent image deviation comparison and calibration.

[0040] The first deviation comparison module generates hue deviation information by comparing the hue deviation between the standard acquired image and the image to be tested. By comparing the hue deviation of the image to be tested, it is possible to detect whether there is hue and color temperature deviation in the train image, and perform hue correction, thereby providing a more accurate image to be tested for subsequent train anomaly detection.

[0041] The second deviation comparison module compares whether there is a sampling deviation between image acquisition devices by using a standard acquisition image to obtain sampling deviation information, and associates the image acquisition device with the sampling deviation information to obtain device deviation information.

[0042] The verification module is used to detect the accuracy of the color tone deviation information, the sampling deviation information and the device deviation information; the verification module can ensure the accuracy and credibility of the comparison result and avoid misjudgment or missed judgment.

[0043] A calibration module receives hue deviation information, sampling deviation information and device deviation information, calls hue calibration parameters corresponding to the hue deviation information, image calibration parameters corresponding to the sampling deviation information and device calibration parameters corresponding to the device deviation information in the cloud database, performs image hue calibration and image position calibration on the image to be tested, and performs hardware calibration on the image acquisition device; the calibration module can calibrate images and image acquisition devices with deviations, thereby eliminating the subsequent impact of sampling deviations on train detection.

[0044] The cloud database is used to store standard captured images, tone calibration parameters, image calibration parameters and device calibration parameters, and can quickly judge and calibrate image deviations, thereby improving image processing efficiency.

[0045] In this embodiment, refer to Figure 1 to Figure 2As shown, the calibration module is provided with a laser ray lamp, which is fixedly connected to the image acquisition device. The laser ray lamp generates a calibration point with a fixed position in the image to be measured. By using the laser ray lamp fixedly connected to the image acquisition device to map the calibration points, the calibration point position can be always at a fixed position in the acquired image. When the position and angle of the image acquisition device are changed, the calibration module is changed together with it, thereby ensuring that the position of the calibration point in the image remains unchanged. The image acquisition devices are evenly distributed on all sides of the train, ensuring all-round image acquisition of the train from multiple angles, and being able to capture all details of the train, providing a rich data source for subsequent image analysis and processing. There are overlapping sampling areas between adjacent image acquisition devices, and the calibration points fall within the overlapping sampling areas on each side of the image acquisition device, ensuring that there are sufficient overlapping parts between images taken at different angles. The overlapping parts can be used in subsequent image deviation comparison and calibration processes to improve the overall consistency and accuracy of the image.

[0046] In this embodiment, refer to Figure 1 to Figure 2 As shown, the first deviation comparison module performs denoising on the image to be tested and compares the hue components of the color space in the image, calculates the hue components of the image to be tested and the hue components of the standard acquired image, reflects the hue difference between the image to be tested and the standard image according to the hue comparison result, generates hue deviation information, and by comparing the hue deviation of the image to be tested, detects whether the train image has hue and color temperature deviations, and performs hue correction, thereby providing a more accurate image to be tested for subsequent train anomaly detection.

[0047] When the first deviation comparison module is working, the image to be tested is subjected to denoising to improve the image quality for subsequent analysis. In this embodiment, a hue comparison algorithm is selected to compare the hue components in the HSV color space of the image. The algorithm is used to identify and quantify the hue shift, local hue change and color temperature change between the two images, and hue deviation information is generated according to the hue comparison results. These changes are usually caused by the light source, which affects the accuracy of subsequent train anomaly detection. Therefore, after the hue of the image to be tested is subjected to deviation comparison, the image color difference is adjusted in time, which can improve the accuracy of train anomaly detection.

[0048] In this embodiment, refer to Figure 1 to Figure 2As shown, the second deviation comparison module selects two adjacent image acquisition devices in turn, obtains the image to be tested and the auxiliary deviation comparison image through the two image acquisition devices respectively, and the second deviation comparison module identifies the coordinates of the calibration points in the acquired image to be tested and the auxiliary deviation comparison image, and calculates the relative position spacing value of the calibration points of the image to be tested and the auxiliary deviation comparison image and the relative position spacing value difference of the calibration points in the standard acquisition image to obtain the calibration point deviation parameter K; the relative position spacing value refers to the pixel distance or actual distance between the corresponding calibration points in the two images, and the standard spacing value is obtained by precise measurement in the system calibration stage, representing the ideal spacing between the calibration points in the images acquired by the adjacent image acquisition devices without deviation. By accurately identifying the coordinates of the calibration points in the image to be tested and comparing the relative position spacing values ​​of the calibration points in the overlapping sampling area in the images acquired by the adjacent image acquisition devices, the possible deviations between the image acquisition devices can be discovered and quantified in time. This method based on calibration points ensures the accuracy and reliability of deviation detection and provides a solid foundation for subsequent calibration work.

[0049] The second deviation comparison module also extracts the overlapping sampling areas in the image to be tested and the auxiliary deviation comparison image acquired by the image acquisition device through image feature matching, and deletes the non-overlapping areas in the image to obtain the overlapping sampling images; the standard acquisition image also includes a standard overlapping image, and the standard overlapping image is an image of the overlapping area of ​​the standard acquisition image and the image acquired by the adjacent image acquisition device; the overlapping deviation parameter S is obtained by comparing the difference in the length of the four sides and the difference in the area between the overlapping sampling image and the standard overlapping image; by comparing the overlapping areas of the image to be tested and the adjacent images with the parameters of the overlapping areas in the standard acquisition image, a new reference factor is provided for the deviation comparison, thereby improving the accuracy of the deviation comparison.

[0050] The second deviation comparison module performs pattern recognition, identifies the marker pattern in the image to be tested, and identifies the matching marker pattern in the standard acquisition image. By comparing the matching degree of the marker pattern in the image to be tested with the marker pattern in the standard acquisition image, the difference between the matching degree and the complete matching degree is calculated to obtain the pattern abnormality deviation parameter M; by performing feature matching on the pattern with obvious image features in the image, accurate comparison factors can be provided for deviation comparison.

[0051] The cloud database also stores a deviation threshold. The second deviation comparison module calculates the sampling deviation value by combining the calibration point deviation parameter K, the overlap deviation parameter S and the pattern abnormality deviation parameter M, and compares the sampling deviation value with the deviation threshold to obtain the sampling deviation information.

[0052] By considering the coordinates of the calibration points, coincident sampling and characteristic pattern factors, the relative positional relationship between image acquisition devices can be more comprehensively evaluated, thereby further improving the comparison accuracy and ensuring the reliability and accuracy of the results. The introduction of the second deviation comparison module makes the calibration process more complete and systematic. It not only focuses on the deviation between adjacent image acquisition devices, but also further verifies the calibration effect through cross-side comparison, providing higher quality data support for subsequent image analysis and processing.

[0053] In this embodiment, refer to Figure 1 to Figure 2 As shown, the calibration point deviation parameter is calculated as follows:

[0054] ;

[0055] ;

[0056] Among them, k1 is the relative position spacing value of the calibration points of the image to be tested and the auxiliary deviation comparison image, (x1, y1) is the coordinate of the calibration point in the image to be tested, (x2, y2) is the coordinate of the calibration point in the auxiliary deviation comparison image, and k0 is the relative position spacing value difference of the calibration points in the standard acquisition image. The smaller K is, the smaller the possibility of deviation.

[0057] In this embodiment, refer to Figure 1 to Figure 2 As shown, the overlap deviation parameter is calculated as follows:

[0058] ;

[0059] Among them, l i is the length of the i-th side of the overlapped sampled image, s is the area of ​​the overlapped sampled image, l0 i is the length of the i-th side of the standard coincidence image, and s0 is the area of ​​the standard coincidence image. By using the coincidence deviation parameter, an accurate comparison factor is provided for the verification of the second deviation comparison module, thereby improving the accuracy of the deviation verification.

[0060] In this embodiment, refer to Figure 1 to Figure 2 As shown, the sampling deviation values ​​are:

[0061] ;

[0062] Among them, Q is the sampling deviation value, α is the coefficient of the calibration point deviation parameter, β is the coefficient of the overlap deviation parameter, and γ is the coefficient of the pattern abnormality deviation parameter; Q0 is the deviation threshold, and the sampling deviation information is obtained by comparing the values ​​of Q and Q0. By setting the deviation threshold, when the sampling deviation value is greater than the deviation threshold, it is judged that the image to be tested has a sampling deviation. By using the deviation threshold, the train sampling deviation comparison accuracy in different application scenarios can be set.

[0063] In this embodiment, refer to Figure 1 to Figure 2 As shown, the verification module verifies the accuracy of the hue deviation information of the image to be tested by comparing the hue deviation between the image to be tested and the adjacent auxiliary deviation comparison image; the image acquisition device also acquires train images from multiple angles, and the verification module detects the accuracy of the sampling deviation information and device deviation information of the image to be tested by feature matching between the image to be tested and the train image. The existence of the verification module ensures the accuracy and reliability of the entire image acquisition and calibration system. By timely verifying the accuracy of the deviation comparison, potential deviation problems can be discovered and corrected in a timely manner, avoiding these deviations from having an adverse effect on subsequent image analysis and processing. At the same time, the verification module also provides an important basis for system maintenance and optimization, which helps to improve the overall performance and stability of the system.

[0064] In this embodiment, refer to Figure 1 to Figure 2 As shown, the calibration module receives the hue deviation information and the sampling deviation information, and matches the corresponding hue calibration parameters and sampling deviation parameters in the cloud database to calibrate the hue and image position of the image to be tested. The first-level deviation value is matched to the corresponding calibration parameters in the cloud database, and the first-level calibration is performed on the acquired image to be tested, and the deviation of the calibrated image to be tested is compared and verified again. The calibration module can ensure the accuracy and reliability of the image acquisition and calibration system. By timely calibrating the hue, color temperature and sampling position of the image, the deviation in the image can be eliminated or reduced, and the quality and consistency of the image can be improved.

[0065] In this embodiment, refer to Figure 1 to Figure 2As shown, the calibration module first receives the hue deviation information and sampling deviation information output by the first deviation comparison module and the second deviation comparison module. After receiving the hue deviation information and sampling deviation information, the calibration module accesses the cloud database and matches the corresponding calibration parameters according to the deviation information. The cloud database stores the calibration parameters corresponding to different deviation values. These parameters are obtained in advance through experiments or calculations and are used to correct the deviation in the image. The matching process searches for the calibration parameters corresponding to the closest deviation information. After matching the calibration parameters, the calibration module will apply these parameters to calibrate the collected image to be tested. The calibration process includes operations such as image hue adjustment, translation, rotation, scaling, and distortion correction, which are intended to eliminate or reduce the deviation in the image. The specific method of calibration depends on the type and degree of deviation, as well as the selected calibration parameters. After completing the first-level calibration, the first calibration module will send the calibrated image to be tested to the deviation comparison module again for deviation comparison. This step is to verify the calibration effect, that is, to check whether there is still deviation in the calibrated image and whether the deviation has been reduced to an acceptable range. At the same time, the verification module will also verify the calibrated image to ensure that the calibration process does not introduce new errors or problems. If there is still a large deviation in the calibrated image, or the verification results show that the calibration effect is not good, the first calibration module may adjust the calibration parameters or take other measures to improve the calibration effect based on the feedback results. In some cases, if the deviation is too large or difficult to completely eliminate through the first-level calibration, the system may trigger a higher-level calibration process or notify maintenance personnel for manual intervention.

[0066] In this embodiment, refer to Figure 1 to Figure 2 As shown, the image acquisition device is provided with a rotation mechanism, the calibration module receives the device deviation information, and matches the corresponding device calibration parameters in the cloud database through the device deviation information, and adjusts the angle and position of the image acquisition device through the device calibration parameters to perform hardware calibration; when the calibration module is in hardware calibration, the calibration module connected to the image acquisition device adjacent to the calibrated image acquisition device emits rays to the target correction position in the overlapping sampling area to form a target calibration point, and adjusts the rotation mechanism of the calibrated image acquisition device so that the calibration point emitted by the calibration module connected to it coincides with the target calibration point to complete the hardware calibration. By using the calibration point to calibrate the device deviation, the accuracy of the calibration can be ensured, and by calibrating the device deviation, the deviation of the image acquisition device can be calibrated in time, ensuring the accuracy of subsequent image acquisition.

[0067] The above shows and describes the basic features, principles and advantages of the present invention. It should be noted that the present invention is not limited to the above embodiments, which are only partial embodiments. Under the premise of not departing from the spirit and scope of the present invention, several improvements and supplements made are considered as the protection scope of the present invention.

Claims

1. A synchronous calibration system for multi-angle train image acquisition, characterized in that: include: An image acquisition module, wherein the image acquisition module includes a plurality of image acquisition devices for acquiring images to be tested; A calibration module, which is fixedly connected to the image acquisition module and is used to map calibration points to the area to be measured within the sampling range of the image acquisition module; A first deviation comparison module, wherein the first deviation comparison module generates hue deviation information by comparing the hue deviation between the standard captured image and the image to be tested; A second deviation comparison module, wherein the second deviation comparison module compares whether there is a sampling deviation between image acquisition devices by using a standard acquisition image to obtain sampling deviation information, and associates the image acquisition device with the sampling deviation information to obtain device deviation information; A verification module, the verification module is used to detect the accuracy of the hue deviation information, the sampling deviation information and the device deviation information; A calibration module, wherein the calibration module receives the hue deviation information, the sampling deviation information and the device deviation information, calls the hue calibration parameters corresponding to the hue deviation information, the image calibration parameters corresponding to the sampling deviation information and the device calibration parameters corresponding to the device deviation information in the cloud database, performs hue calibration and image position calibration on the image to be tested, and performs hardware calibration on the image acquisition device; A cloud database, wherein the cloud database is used to store standard captured images, tone calibration parameters, image calibration parameters, and device calibration parameters; The second deviation comparison module selects two adjacent image acquisition devices in sequence, and acquires the image to be tested and the auxiliary deviation comparison image through the two image acquisition devices respectively. The second deviation comparison module identifies the coordinates of the calibration points in the acquired image to be tested and the auxiliary deviation comparison image, and calculates the difference between the relative position spacing value of the calibration points in the image to be tested and the auxiliary deviation comparison image and the relative position spacing value of the calibration points in the standard acquisition image, and obtains the calibration point deviation parameter K; The second deviation comparison module also extracts the overlapping sampling areas in the image to be tested and the auxiliary deviation comparison image collected by the image acquisition device through image feature matching, and deletes the non-overlapping areas in the image to obtain the overlapping sampling images; the standard acquisition image also includes a standard overlapping image, which is an image of the overlapping area of ​​the standard acquisition image and the image collected by the adjacent image acquisition device; the overlap deviation parameter S is obtained by comparing the difference in the length of the four sides and the difference in the area between the overlapping sampling image and the standard overlapping image; The second deviation comparison module performs pattern recognition, identifies the marker pattern in the image to be tested, and identifies the matching marker pattern in the standard acquisition image, and calculates the difference between the matching degree and the complete matching degree by comparing the matching degree between the marker pattern in the image to be tested and the marker pattern in the standard acquisition image, thereby obtaining the pattern abnormality deviation parameter M; The cloud database also stores a deviation threshold. The second deviation comparison module calculates the sampling deviation value by combining the calibration point deviation parameter K, the overlap deviation parameter S and the pattern abnormality deviation parameter M, and compares the sampling deviation value with the deviation threshold to obtain the sampling deviation information.

2. The synchronous calibration system for multi-angle train image acquisition according to claim 1, characterized in that: The calibration module is provided with a laser ray lamp, which is fixedly connected to the image acquisition device. The laser ray lamp generates calibration points with fixed positions in the image to be measured. The image acquisition devices are evenly distributed on both sides of the train. There are overlapping sampling areas between adjacent image acquisition devices. The calibration points fall within the overlapping sampling areas on each side of the image acquisition device.

3. The synchronous calibration system for multi-angle train image acquisition according to claim 1, characterized in that: The first deviation comparison module performs denoising on the image to be tested and compares the hue components of the color space in the image, calculates the hue components of the image to be tested and the hue components of the standard captured image, reflects the hue difference between the image to be tested and the standard image according to the hue comparison result, and generates hue deviation information.

4. The synchronous calibration system for multi-angle train image acquisition according to claim 1, characterized in that: The calibration point deviation parameter calculation is: ； ； Among them, k1 is the relative position spacing value of the calibration points in the image to be tested and the auxiliary deviation comparison image, (x1, y1) is the coordinates of the calibration points in the image to be tested, (x2, y2) is the coordinates of the calibration points in the auxiliary deviation comparison image, and k0 is the difference in the relative position spacing value of the calibration points in the standard acquisition image.

5. The synchronous calibration system for multi-angle train image acquisition according to claim 1, characterized in that: The overlap deviation parameter calculation is: ； Among them, l i is the length of the i-th side of the overlapped sampled image, s is the area of ​​the overlapped sampled image, l0 i is the length of the i-th side of the standard coincidence image, and s0 is the area of ​​the standard coincidence image.

6. The synchronous calibration system for multi-angle train acquisition images according to claim 1, characterized in that: The sampling deviation values ​​are: ； Among them, Q is the sampling deviation value, α is the coefficient of the calibration point deviation parameter, β is the coefficient of the overlap deviation parameter, and γ is the coefficient of the pattern abnormality deviation parameter; Q0 is the deviation threshold, and the sampling deviation information is obtained by comparing the numerical values ​​of Q and Q0.

7. The synchronous calibration system for multi-angle train image acquisition according to claim 1, characterized in that: The verification module verifies the accuracy of the hue deviation information of the image to be tested by comparing the hue deviation between the image to be tested and the adjacent auxiliary deviation comparison image; the image acquisition device also acquires train images from multiple angles, and the verification module detects the accuracy of the sampling deviation information and the equipment deviation information of the image to be tested by performing feature matching between the image to be tested and the train image.

8. The synchronous calibration system for multi-angle train image acquisition according to claim 1, characterized in that: The calibration module receives the hue deviation information and the sampling deviation information, matches the corresponding hue calibration parameters and sampling deviation parameters in the cloud database, and calibrates the hue and image position of the image to be measured.

9. The synchronous calibration system for multi-angle train image acquisition according to claim 1, characterized in that: The image acquisition device is provided with a rotation mechanism, the calibration module receives device deviation information, matches the corresponding device calibration parameters in the cloud database through the device deviation information, adjusts the angle and position of the image acquisition device through the device calibration parameters, and performs hardware calibration; when the calibration module performs hardware calibration, the calibration module connected to the image acquisition device adjacent to the calibrated image acquisition device emits rays to the target correction position in the overlapping sampling area to form a target calibration point, and adjusts the rotation mechanism of the calibrated image acquisition device so that the calibration point emitted by the calibration module connected to it coincides with the target calibration point, thereby completing the hardware calibration.

Images