Track displacement high-precision detection method and system, medium and electronic equipment

Abstract

The invention provides a track displacement high-precision detection method and system, a medium and electronic equipment, and relates to the technical field of track measurement, and the method comprises the steps: obtaining a measurement image and a reference image in real time, and representing the measurement image as an image containing a measurement target configured on a track, the reference image is represented as an image including a reference target configured on a preset reference platform, and the preset reference platform is located on one side of a track at the same position as the measurement target; determining a first displacement amount of the measurement target according to the measurement images obtained at the preset historical moment and the current target moment, and determining a second displacement amount of the reference target according to the reference images obtained at the preset historical moment and the current target moment; and correcting the first displacement through the second displacement to obtain the target displacement of the track. The problem that in the prior art, track displacement detection equipment is inconvenient to use is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of track measurement, and particularly to a high-precision detection method, system, medium and electronic device for track displacement. Background Art

[0002] In recent years, the railway track displacement detection technology has become an important research direction for ensuring train running safety and reducing operation and maintenance costs. The existing technologies mainly monitor parameters such as track lateral offset, longitudinal settlement and turnout deformation through manual inspection or contact sensors.

[0003] Although the traditional manual inspection method does not require complex equipment investment, it has obvious deficiencies such as low efficiency, large subjective errors and inability to achieve continuous online monitoring; To solve the problem of low efficiency of the above-mentioned manual inspection method, the contact sensors in the existing technologies can achieve real-time monitoring without manual operation. However, since the above-mentioned contact sensors need to be installed on the track, and the above-mentioned contact sensors are highly sensitive to vibration and have poor stability under harsh climate conditions, it is difficult for them to adapt to complex railway operating environments and is inconvenient to use. Summary of the Invention

[0004] Aiming at the deficiencies in the existing technologies, the present invention provides a high-precision detection method, system, medium and electronic device for track displacement, and solves the problem that the existing track displacement detection devices are inconvenient to use.

[0005] The present invention provides a high-precision detection method for track displacement, including: Real-time obtaining a measurement image and a reference image, where the measurement image is an image including a measurement target configured on the track, and the reference image is an image including a reference target configured on a preset reference platform, and the preset reference platform is located on one side of the track at the same position as the measurement target; Determining a first displacement amount of the measurement target according to the measurement image obtained at a preset historical moment and a current target moment, and determining a second displacement amount of the reference target according to the reference image obtained at the preset historical moment and the current target moment; Correcting the first displacement amount by the second displacement amount to obtain a target displacement amount of the track.

[0006] The technical solution provided by the present invention has at least the following beneficial effects: By obtaining images of the measurement target and the reference target, and determining the target displacement of the track based on the first displacement and the second displacement of the reference target and the measurement target at a preset historical moment and the current target moment, this non-contact visual measurement method avoids installing contact physical sensors on the track, reduces the interference to the track structure and the complexity of equipment installation. At the same time, through the setting of the reference target, it can automatically cancel out the slight jitter that may be generated by the camera or the installation base, ensuring the final detection accuracy without manual intervention, being more convenient to use, and also having significant advantages such as low cost, light equipment, and no need for high-precision on-site calibration, greatly reducing the deployment and operation and maintenance costs.

[0007] In a high-precision track displacement detection method provided by the present invention, determining the first displacement of the measurement target based on the measurement images obtained at the preset historical moment and the current target moment includes: Based on the initial measurement image obtained at the preset historical moment, determining the initial structural feature image of the measurement target in the initial measurement image, and determining the initial measurement coordinates of the pixels at the center position of the measurement target in the initial structural feature image; Based on the real-time measurement image obtained at the current target moment, determining the real-time structural feature image of the measurement target in the real-time measurement image, and determining the real-time measurement coordinates of the pixels at the center position of the measurement target in the real-time structural feature image; Determining the offset between the initial measurement coordinates and the real-time measurement coordinates, and combining with a preset mapping coefficient to determine the first displacement of the measurement target at the current target moment, where the mapping coefficient represents the conversion relationship between the pixel range occupied by the imaging of the target object on the camera's photosensitive device and its physical size.

[0008] The technical solution provided by the present invention has at least the following beneficial effects: By extracting the initial structural feature image and the real-time structural feature image to determine the initial measurement coordinates and the real-time measurement coordinates at the center position of the measurement target, the extraction accuracy of the above initial measurement coordinates and real-time measurement coordinates can be improved, thereby improving the detection accuracy of this method.

[0009] In a high-precision track displacement detection method provided by the present invention, determining the real-time structural feature image of the measurement target in the real-time measurement image based on the real-time measurement image obtained at the current target moment includes: Obtaining the historical measurement coordinates of the pixels at the center position obtained at the previous moment of the current target moment, and based on the historical measurement coordinates, determining the ROI region of the real-time measurement image, where the ROI region is the region covered by the measurement target; Zero the pixel values of the regions outside the ROI region in the real-time measurement image, and perform dynamic threshold binarization on the processed real-time measurement image to obtain a target measurement image; Based on the Hough circle transform, identify and extract the structural features of the measurement target in the target measurement image to obtain a real-time structural feature image of the measurement target.

[0010] The technical solution provided by the present invention has at least the following beneficial effects: Based on the division of the dynamic ROI region, regions can be set separately for each measurement target and local image processing can be performed, effectively reducing the ineffective calculations for the entire image, improving the system response speed and energy efficiency ratio, while significantly reducing the interference of image background noise and enhancing the accuracy of feature positioning.

[0011] In a high-precision track displacement detection method provided by the present invention, the step of determining the offset between the initial measurement coordinates and the real-time measurement coordinates, and combining with a preset mapping coefficient to determine the first displacement amount of the measurement target at the current target moment includes: According to the horizontal coordinate offset and the vertical coordinate offset between the initial measurement coordinates and the real-time measurement coordinates, and combining with a preset mapping coefficient, determine the horizontal displacement and the vertical displacement of the measurement target, and use the horizontal displacement and the vertical displacement as the first displacement amount of the measurement target at the current target moment.

[0012] In a high-precision track displacement detection method provided by the present invention, the step of determining the mapping coefficient includes: Perform sub-pixel accuracy correction on the pixels at the edge position of the measurement target in the initial structural feature image; Based on the corrected initial structural feature map, determine the pixel radius of the measurement target in the initial structural feature image; Obtain the true radius of the measurement target, and use the ratio of the true radius to the pixel radius as the mapping coefficient.

[0013] In a high-precision track displacement detection method provided by the present invention, the step of performing sub-pixel accuracy correction on the pixels at the edge position of the measurement target in the initial structural feature image; based on the corrected initial structural feature image, determine the pixel radius of the measurement target in the initial structural feature image includes: Based on the edge difference method, the gray center of gravity method, and the fitting optimization method, perform sub-pixel accuracy correction on the pixels at the edge position of the measurement target in the initial structural feature image respectively; Based on the gradient consistency algorithm, according to each corrected initial structural feature image, determine the edge smoothness and consistency degree of each initial structural feature image respectively; Take the initial structural feature image with the highest degree of edge smoothness and consistency as the target initial structural feature image, and determine the pixel radius of the measurement target according to the target initial structural feature image.

[0014] The technical solution provided by the present invention has at least the following beneficial effects: By introducing a visual measurement algorithm with sub-pixel accuracy correction, on the basis of traditional pixel-level positioning, the edge difference method, the gray center of gravity method, and the fitting optimization method are further used to improve the accuracy of the target pixels at the target position to the sub-pixel level. Combined with the conversion of the physical size of the measurement target, millimeter or sub-millimeter level displacement measurement accuracy is achieved, meeting the requirements for monitoring the small deformation of the track structure, and is superior to the traditional vision system.

[0015] In a high-precision track displacement detection method provided by the present invention, determining the real-time measurement coordinates of the pixels at the center position of the measurement target in the real-time structural feature image includes: Based on the edge difference method, the gray center of gravity method, and the fitting optimization method, perform sub-pixel accuracy correction on the pixels at the edge position of the measurement target in the real-time structural feature image respectively; Based on the gradient consistency algorithm, determine the edge smoothness and consistency degree of each real-time structural feature image respectively according to the corrected real-time structural feature images; Take the real-time structural feature image with the highest degree of edge smoothness and consistency as the target real-time structural feature image, and determine the real-time measurement coordinates of the pixels at the center position of the measurement target according to the target real-time structural feature image.

[0016] The present invention also provides a high-precision track displacement detection system, including: An image acquisition module that real-time acquires a measurement image and a reference image. The measurement image is an image containing a measurement target configured on the track, and the reference image is an image containing a reference target configured on a preset reference platform. The preset reference platform is located on one side of the track at the same position as the measurement target; An image processing module that determines the first displacement amount of the measurement target according to the measurement image obtained at a preset historical moment and the current target moment, and determines the second displacement amount of the reference target according to the reference image obtained at the preset historical moment and the current target moment; A displacement correction module that corrects the first displacement amount through the second displacement amount to obtain the target displacement amount of the track.

[0017] The present invention also provides a computer-readable storage medium storing instructions that, when run on a terminal device, cause the terminal device to execute a high-precision track displacement detection method as described above.

[0018] The present invention also provides an electronic device, including a memory, a processor, and a program stored on the memory and running on the processor, where the processor executes the program to implement a high-precision track displacement detection method as described above. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 is a schematic flowchart of a high-precision track displacement detection method provided by the present invention; Figure 2 is a processing logic diagram of a high-precision track displacement detection method provided by the present invention; Figure 3 is a schematic flowchart of determining the center pixel coordinates or radius in a high-precision track displacement detection method provided by the present invention; Figure 4 is a schematic diagram of the dynamic ROI area marking result; Figure 5 is a schematic diagram of the track displacement detection result; Figure 6 is a schematic structural diagram of a high-precision track displacement detection system provided by the present invention; Figure 7 is a schematic structural diagram of the electronic device provided by the present invention.

[0020] In the drawings, the list of components represented by each reference numeral is as follows: 10. Electronic device, 11. Processor, 12. Read-only memory (ROM), 13. Random access memory (RAM), 14. Bus, 15. Input / output (I / O) interface, 16. Input unit, 17. Output unit, 18. Storage unit, 19. Communication unit. DETAILED DESCRIPTION

[0021] The principles and features of the present invention are described below, and the examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0022] The present invention provides a high-precision track displacement detection method. Please refer to Figure 1 as shown, including: Obtain a measurement image and a reference image in real time. The measurement image is an image containing a measurement target configured on a track, and the reference image is an image containing a reference target configured on a preset reference platform. The preset reference platform is located on one side of the track at the same position as the measurement target; Determine the first displacement of the measurement target based on the measurement image obtained at the preset historical moment and the current target moment, and determine the second displacement of the reference target based on the reference image obtained at the preset historical moment and the current target moment; Correct the first displacement by the second displacement to obtain the target displacement of the track.

[0023] By acquiring the images of the measurement target and the reference target, and based on the first displacement and the second displacement of the reference target and the measurement target at the preset historical moment and the current target moment, to determine the target displacement of the track. Through this non-contact visual measurement method, it is possible to avoid installing contact physical sensors on the track, reduce the interference to the track structure and the complexity of equipment installation. At the same time, through the setting of the reference target, it is possible to automatically cancel out the small jitters that may be generated by the camera or the installation base, ensuring the final detection accuracy without manual intervention, which is more convenient to use. At the same time, it also has significant advantages such as low cost, light equipment, and no need for high-precision on-site calibration, greatly reducing the deployment and operation and maintenance costs.

[0024] It should be noted that the monitoring device for acquiring the measurement image and the reference image in real time is the same camera to avoid errors. In this embodiment, the above camera is an infrared night vision camera, which is equipped with an infrared emission device to ensure that the measurement target and the reference target reflect infrared light to form images. At the same time, both the measurement target and the reference target are configured with high-reflectivity materials (such as white reflective materials), that is, both the measurement target and the reference target are passive structures and do not emit light by themselves. At the same time, the shapes of the measurement target and the reference target are optimized to an infrared type structure composed of a combination of curves and straight lines to enhance the feature robustness.

[0025] Optionally, to improve the detection accuracy of the rail displacement, determining the first displacement of the measurement target based on the measurement image obtained at the preset historical moment and the current target moment includes: Based on the initial measurement image obtained at the preset historical moment, determine the initial structural feature image of the measurement target in the initial measurement image; determine the initial measurement coordinates of the pixels at the center position of the measurement target according to the initial structural feature image. In this embodiment, the above measurement target and the reference target have the same shape and are both circular target structures; Based on the real-time measurement image obtained at the current target moment, determine the real-time structural feature image of the measurement target in the real-time measurement image; determine the real-time measurement coordinates of the pixels at the center position of the measurement target according to the real-time structural feature image; Determine the offset between the initial measurement coordinates and the real-time measurement coordinates, and combine with the preset mapping coefficient to determine the first displacement of the measurement target at the current target moment. The mapping coefficient represents the conversion relationship between the pixel range occupied by the imaging of the target object on the camera's photosensitive device and its physical size.

[0026] Similarly, the step of determining the second displacement of the reference target at the preset historical moment and the current target moment is the same as that of determining the first displacement of the measurement target, that is: Based on the initial reference image obtained at the preset historical moment, determine the initial reference structure feature image of the reference target in the initial reference image; determine the initial reference coordinates of the pixels at the center position of the reference target according to the initial reference structure feature image; Based on the real-time reference image obtained at the current target moment, determine the real-time reference structure feature image of the reference target in the real-time reference image; determine the real-time reference coordinates of the pixels at the center position of the reference target according to the real-time reference structure feature image; Determine the offset between the initial reference coordinates and the real-time reference coordinates, and combine the preset mapping coefficient to determine the second displacement of the reference target at the current target moment.

[0027] Specifically, based on the real-time measurement image obtained at the current target moment, determine the real-time structure feature image of the measurement target in the real-time measurement image, including: Set independent ROI regions respectively through dynamic ROI coding. Use the position of the target in the previous moment image (i.e., the measurement coordinates or reference coordinates) to perform predictive ROI division on the image of the next moment. The ROI region is the region covered by the measurement target or the reference target; Based on this, obtain the historical measurement coordinates of the pixels at the center position of the measurement target obtained at the previous moment of the current target moment. Based on the historical measurement coordinates, determine the ROI1 region of the real-time measurement image. The ROI1 region is the region covered by the measurement target; Similarly, obtain the historical reference coordinates of the pixels at the center position of the reference target obtained at the previous moment of the current target moment. Based on the historical reference coordinates, determine the ROI2 region of the real-time reference image. The ROI2 region is the region covered by the reference target; Subsequently, set the pixel values of the regions outside the ROI regions in the measurement image and the reference image to zero, and perform dynamic threshold binarization on the processed measurement image and reference image. The dynamic threshold is adaptively adjusted according to the infrared image characteristics, so as to obtain the target measurement image and the target reference image at the current target moment; Based on the Hough circle transform, identify and extract the structural features of the measurement target in the target measurement image, identify the circular structures of the measurement target and the reference target, and obtain the real-time structure feature image of the measurement target in the real-time measurement image and the real-time reference structure feature image of the reference target in the real-time reference image, so as to facilitate the determination of the corresponding coordinates of the center pixels of the measurement target and the reference target.

[0028] Based on the division of the dynamic ROI region, regions can be set separately for each measurement target and reference target, and local image processing can be performed, effectively reducing the ineffective calculations for the entire image, improving the system response speed and energy efficiency ratio. At the same time, the interference of the image background noise is greatly reduced, and the accuracy of feature positioning is enhanced.

[0029] Specifically, determine the offset between the initial measurement coordinates and the real-time measurement coordinates, and combine with the preset mapping coefficient to determine the first displacement of the measurement target at the current target moment, including: According to the horizontal offset and vertical offset between the initial measurement coordinates and the real-time measurement coordinates, and combine with the preset mapping coefficient to determine the horizontal displacement and vertical displacement of the measurement target, and use the horizontal displacement and vertical displacement as the first displacement of the measurement target. Similarly, according to the determined offset between the initial reference coordinates and the real-time reference coordinates, the horizontal displacement and vertical displacement of the reference target can be obtained as the second displacement of the reference target; Optionally, the steps for determining the mapping coefficient include: Perform sub-pixel precision correction on the pixels at the edge position of the measurement target in the initial structural feature image; Based on the corrected initial structural feature map, determine the pixel radius of the measurement target in the initial structural feature image; Obtain the real radius of the measurement target, and use the ratio of the real radius to the pixel radius as the mapping coefficient.

[0030] Specifically, to improve the accuracy of the mapping coefficient, the above-mentioned sub-pixel precision correction of the pixels at the edge position of the measurement target in the initial structural feature image; based on the corrected initial structural feature image, determine the pixel radius of the measurement target in the initial structural feature image, specifically including: Based on the edge difference method, gray center of gravity method, and fitting optimization method, perform sub-pixel precision correction on the pixels at the edge position of the measurement target in the initial structural feature image respectively; Based on the gradient consistency algorithm, according to the corrected initial structural feature images, determine the corresponding consistency scores of the respective initial structural feature images. The consistency score represents the edge smoothness and consistency degree of the circular structure. Among them, the lower the consistency score, the higher the edge smoothness and consistency degree of the circular structure; Use the initial structural feature image with the highest edge smoothness and consistency degree (i.e., the lowest consistency score) as the target initial structural feature image, and determine the pixel radius of the measurement target according to the target initial structural feature image.

[0031] By introducing a visual measurement algorithm with sub-pixel precision correction, based on the traditional pixel-level positioning, the edge difference method, the gray centroid method, and the fitting optimization method are further used to improve the precision of the target pixels at the target position to the sub-pixel level. Combining with the conversion of the physical size of the measurement target, the displacement measurement precision of millimeter or sub-millimeter level is achieved, meeting the requirements of monitoring the tiny deformation of the track structure, and being superior to the traditional vision system.

[0032] Furthermore, to improve the detection precision, when obtaining the initial reference coordinates, the initial measurement coordinates, the real-time reference coordinates, and the real-time measurement coordinates, the sub-pixel precision correction is carried out by the above-mentioned edge difference method, the gray centroid method, and the fitting optimization method, and the consistency score is calculated to determine. Since the calculation methods of the initial reference coordinates, the initial measurement coordinates, the real-time reference coordinates, and the real-time measurement coordinates are the same, here, taking the calculation of the real-time measurement coordinates as an example for illustrative explanation: That is, the real-time measurement coordinates of the pixels at the center position of the measurement target in the real-time structure feature image are determined, including: Based on the edge difference method, the gray centroid method, and the fitting optimization method, the sub-pixel precision correction is respectively carried out on the pixels at the edge position of the measurement target in the real-time structure feature image; Based on the gradient consistency algorithm, according to each corrected real-time structure feature image, the edge smoothness and consistency degree of each real-time structure feature image are respectively determined; The real-time structure feature image with the highest edge smoothness and consistency degree is used as the target real-time structure feature image, and according to the target real-time structure feature image, the real-time measurement coordinates of the pixels at the center position of the measurement target are determined.

[0033] In summary, in a more specific embodiment provided by the present invention, please refer to Figure 2 as shown, including the following steps: 1. Install an active infrared night vision camera at the monitoring end, equipped with an infrared emission device, to ensure that the target reflects infrared light to form an image. Set a reference target made of a high-reflectivity material at the measurement end, and install measurement targets with the same size on the railway track. The shape of the target is optimized to an infrared type structure combined with curves and straight lines to enhance the feature robustness.

[0034] 2. Before image acquisition, set independent ROI regions through dynamic ROI respectively. And distinguish the targets by position coordinate coding (such as ROI1 is the reference target, ROI2 is the measurement target); Meanwhile, set independent ROI regions through dynamic ROI coding respectively. Using the position of the target in the previous moment image (i.e., the measurement coordinate or the reference coordinate), the predictive ROI division is carried out on the image in the next moment, and the ROI region is the area covered by the measurement target; 3. Please combine here Figure 3As shown, after collecting the initial reference image, real-time reference image, initial measurement image, and real-time measurement image, extract the preset ROI1 regions in the initial reference image and real-time reference image, and the preset ROI2 regions in the initial measurement image and real-time measurement image. The ROI1 region and ROI2 region at a certain moment are used as references. Figure 4 As shown, and perform pixel value zeroing (i.e., blackening the image) on the regions outside the ROI to exclude interference from irrelevant backgrounds. Stitch and reconstruct the ROI regions into the original image framework, and perform dynamic threshold binarization on each processed image. This dynamic threshold is adaptively adjusted according to the characteristics of the infrared image to enhance the target edge features, and obtain the initial structure feature image corresponding to the initial measurement image, the real-time structure feature image corresponding to the real-time measurement image, the initial reference structure feature image corresponding to the initial reference image, and the real-time reference structure feature image corresponding to the real-time reference image; 4. According to the obtained various structure feature images, use edge interpolation, gray centroid method, and fitting optimization method to perform sub-pixel accuracy correction on the pixels at the center position of the measurement target or reference target therein. Based on the gradient consistency algorithm, according to the corrected various structure feature images, respectively determine the edge smoothness and consistency degree of each structure feature image, and take the structure feature image with the highest edge smoothness and consistency degree as the target structure feature image, so as to output the initial measurement coordinates of the measurement target center pixel corresponding to the target initial structure feature image, the real-time measurement coordinates of the measurement target center pixel corresponding to the target real-time structure feature image, the initial reference coordinates of the reference target center pixel corresponding to the target initial reference structure feature image, and the real-time reference coordinates of the reference target center pixel corresponding to the target real-time reference structure feature image; 5. On the basis of circular extraction, establish a pixel-physical mapping model based on the conversion relationship between the pixel range occupied by the target object in the imaging on the camera photosensitive device and its physical size to obtain the mapping coefficient. It includes: using edge interpolation, gray centroid method, and fitting optimization method to perform sub-pixel accuracy correction on the target pixels at the edge contour position of the measurement target in the initial structure feature image. Based on the gradient consistency algorithm, according to the corrected various initial structure feature images, respectively determine the edge smoothness and consistency degree of each initial structure feature image, and take the initial structure feature image with the highest edge smoothness and consistency degree as the target initial structure feature image. According to the target initial structure feature image, determine the pixel radius of the measurement target, and further improve the center location accuracy to 0.01 pixel level; Compare the true radius r_real of the measurement target with the corrected pixel radius r_pixel to obtain the actual physical distance d_pixel corresponding to one pixel point. The calculation formula is: d_pixel = r_real / r_pixel. Take the actual physical distance d_pixel as the mapping coefficient; Meanwhile, take the obtained initial reference coordinates and initial measurement coordinates as the standard position coordinates when the reference target and the measurement target have not undergone any displacement. 6. According to the offsets of the initial reference coordinates and the real-time reference coordinates, and the offsets of the initial measurement coordinates and the real-time measurement coordinates, use the following formula to calculate the displacement: Dispx = |Δx| × d_pixel; Dispy = |Δy| × d_pixel; Where, Δx represents the abscissa offset of the initial reference coordinates and the real-time reference coordinates (or the initial measurement coordinates and the real-time measurement coordinates) in the X-axis direction, and Δy represents the ordinate offset of the initial reference coordinates and the real-time reference coordinates (or the initial measurement coordinates and the real-time measurement coordinates) in the Y-axis direction. Dispx and Dispy are the true physical displacements of the reference target (or the measurement target) in the corresponding horizontal or vertical directions. For easy distinction, Dispx1 and Dispy1 respectively represent the true physical displacements of the measurement target in the horizontal and vertical directions, and Dispx2 and Dispy2 respectively represent the true physical displacements of the reference target in the horizontal and vertical directions.

[0035] 7. To further improve the measurement accuracy and system stability, the system introduces a reference target offset correction mechanism. Considering the influence of slight movement or jitter of the reference target, a reference drift correction mechanism is introduced.

[0036] At this time, the lateral displacement ΔDispx of the track after this drift correction is: ΔDispx = |Dispx1 - Dispx2|; The longitudinal displacement ΔDispy of the track after this drift correction is: ΔDispy = |Dispy1 - Dispy2|; Finally, take the obtained lateral displacement ΔDispx of the track and the longitudinal displacement ΔDispy of the track as the target displacement of the track. The display reference of the final target displacement is Figure 5 as shown.

[0037] After adopting the above method, through the non-contact visual measurement method of the infrared camera and the target, it is avoided to install contact physical sensors on the track, reducing the interference to the track structure and the equipment installation complexity. Compared with the millimeter-wave radar and laser ranging solutions, the infrared camera system has the advantages of low cost, light equipment, and no need for high-precision on-site calibration, significantly reducing the deployment and operation and maintenance costs.

[0038] The present invention adopts active infrared night vision imaging technology, which can stably acquire images in harsh environments such as at night, in tunnels, in rain and fog, and in dust. Its anti-interference ability is stronger than that of visible light camera systems that rely on natural light. Combined with the design of a high-reflectivity target, even under conditions of track vibration or strong background interference, stable target recognition and tracking can still be achieved through a dynamic ROI division mechanism and a structural feature extraction algorithm.

[0039] The target adopted by the present invention has a completely passive structure, does not contain a power supply, electronic components or an active emission device, and only relies on passive reflection of infrared light for imaging. It has a simple structure, a long service life, and extremely low maintenance costs. It has higher reliability and environmental adaptability, and is especially suitable for long-term field deployment and large-scale track network layout.

[0040] Based on the dynamic ROI division and position coding mechanism, the system of the present invention can separately set regions for each target and perform local image processing, effectively reducing the invalid calculations for the entire image, improving the system response speed and energy efficiency ratio. At the same time, it greatly reduces the interference of image background noise and enhances the accuracy of feature positioning.

[0041] By introducing a sub-pixel-level visual measurement algorithm, on the basis of traditional pixel-level positioning, the present invention further uses fitting and interpolation techniques to improve the accuracy of the center coordinate of the circle to the sub-pixel level. Combined with the conversion of the physical size of the target, millimeter or sub-millimeter-level displacement measurement accuracy is achieved, meeting the requirements for monitoring the tiny deformation of the track structure, which is superior to traditional vision systems.

[0042] The present invention supports a displacement compensation mechanism for the reference target. By calculating the difference to compare the directional offset between the measurement target and the reference target, it automatically cancels the possible small jitters of the camera or the installation base, without the need for manual intervention and correction, further improving the stability and accuracy of the system in dynamic scenarios.

[0043] The present invention does not rely on complex deep learning models or high-power GPU processors. The core algorithm is based on classical image processing and geometric modeling, and can be directly deployed on edge devices or low-power embedded platforms, effectively reducing energy consumption and extending the continuous working time of the system under field conditions, meeting the requirements for long-term unattended track inspection.

[0044] The present invention also provides a high-precision track displacement detection system. Please refer here Figure 6 as shown, including: An image acquisition module that acquires measurement images and reference images in real time. The measurement image is characterized as an image containing a measurement target configured on the track, and the reference image is characterized as an image containing a reference target configured on a preset reference platform. The preset reference platform is located on one side of the track at the same position as the measurement target; The image processing module determines the first displacement of the measurement target according to the measurement image obtained at the preset historical moment and the current target moment, and determines the second displacement of the reference target according to the reference image obtained at the preset historical moment and the current target moment; The displacement correction module corrects the first displacement by the second displacement to obtain the target displacement of the track.

[0045] In a high-precision track displacement detection system provided by the present invention, the image processing module includes: The feature extraction unit determines the initial structural feature image of the measurement target in the initial measurement image based on the initial measurement image obtained at the preset historical moment, and determines the real-time structural feature image of the measurement target in the real-time measurement image based on the real-time measurement image obtained at the current target moment; The center fitting unit determines the initial measurement coordinates of the pixels at the center position of the measurement target according to the initial structural feature image, and determines the real-time measurement coordinates of the pixels at the center position of the measurement target according to the real-time structural feature image; The displacement calculation unit determines the offset between the initial measurement coordinates and the real-time measurement coordinates, and combines the preset mapping coefficient to determine the first displacement of the measurement target at the current target moment. The mapping coefficient represents the conversion relationship between the pixel range occupied by the imaging of the target object on the camera photosensitive device and its physical size.

[0046] Further, the feature extraction unit specifically includes: Obtain the historical measurement coordinates of the pixels at the center position obtained at the previous moment of the current target moment, and determine the ROI region of the real-time measurement image based on the historical measurement coordinates. The ROI region is the region covered by the measurement target; Set the pixel values of the regions outside the ROI region in the real-time measurement image to zero, and perform dynamic threshold binarization on the processed real-time measurement image to obtain the target measurement image; Based on the Hough circle transform, identify and extract the structural features of the measurement target in the target measurement image to obtain the real-time structural feature image of the measurement target.

[0047] Further, the displacement calculation unit specifically includes: According to the horizontal offset and vertical offset between the initial measurement coordinates and the real-time measurement coordinates, and combining the preset mapping coefficient, determine the horizontal displacement and vertical displacement of the measurement target, and use the horizontal displacement and vertical displacement as the first displacement of the measurement target at the current target moment.

[0048] Further, the steps for determining the mapping coefficient in the displacement calculation unit include: Perform sub-pixel precision correction on the pixels at the edge position of the measurement target in the initial structural feature image; Based on the corrected initial structural feature map, determine the pixel radius of the measurement target in the initial structural feature image; Obtain the true radius of the measurement target, and use the ratio of the true radius to the pixel radius as the mapping coefficient.

[0049] Further, the displacement calculation unit further includes: Based on the edge difference method, the gray center of gravity method, and the fitting optimization method, respectively perform sub-pixel accuracy correction on the pixels at the edge position of the measurement target in the initial structural feature image; Based on the gradient consistency algorithm, according to each corrected initial structural feature image, respectively determine the edge smoothness and consistency degree of each initial structural feature image; Take the initial structural feature image with the highest edge smoothness and consistency degree as the target initial structural feature image, and determine the pixel radius of the measurement target according to the target initial structural feature image.

[0050] Further, the center fitting unit includes: Based on the edge difference method, the gray center of gravity method, and the fitting optimization method, respectively perform sub-pixel accuracy correction on the pixels at the edge position of the measurement target in the real-time structural feature image; Based on the gradient consistency algorithm, according to each corrected real-time structural feature image, respectively determine the edge smoothness and consistency degree of each real-time structural feature image; Take the real-time structural feature image with the highest edge smoothness and consistency degree as the target real-time structural feature image, and determine the real-time measurement coordinates of the pixels at the center position of the measurement target according to the target real-time structural feature image.

[0051] The present invention also provides a computer-readable storage medium, in which instructions are stored. When the instructions run on a terminal device, the terminal device is enabled to execute a high-precision detection method for track displacement as described above.

[0052] The present invention also provides an electronic device, including a memory, a processor, and a program stored on the memory and running on the processor. When the processor executes the program, it implements a high-precision detection method for track displacement as described above.

[0053] Figure 7FIG. shows a schematic structural diagram of an electronic device 10 that can be used to implement the present invention. The electronic device is intended to represent various forms of digital computers, such as, laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as, personal digital processors, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are only examples and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0054] As Figure 7 shown, the electronic device 10 includes at least one processor 11, and a memory communicatively connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc. The memory stores a computer program executable by the at least one processor. The processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 into the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0055] Multiple components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a magnetic disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0056] The processor 11 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the processor 11 include but are not limited to a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as a high-precision detection method for track displacement.

[0057] In some embodiments, a high-precision detection method for track displacement can be implemented as a computer program tangibly embodied in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed onto the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the high-precision detection method for track displacement described above can be performed. Alternatively, in other embodiments, the processor 11 can be configured to execute a high-precision detection method for track displacement by any other suitable means (e.g., by means of firmware).

[0058] The various embodiments of the systems and techniques described above in this document can be implemented in digital electronic circuitry, integrated circuit systems, field-programmable gate arrays (FPGA), application-specific integrated circuits (ASIC), application-specific standard products (ASSP), systems-on-a-chip (SOC), complex programmable logic devices (CPLD), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special-purpose or general-purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit the data and instructions to the storage system, the at least one input device, and the at least one output device.

[0059] The computer programs for implementing the methods of the present invention can be written in any combination of one or more programming languages. These computer programs can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus, such that when the computer programs are executed by the processor, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The computer programs can be executed entirely on the machine, partially on the machine, as a stand-alone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0060] In the context of the present invention, a computer-readable storage medium can be a tangible medium that can contain or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. The computer-readable storage medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. Alternatively, the computer-readable storage medium can be a machine-readable signal medium. More specific examples of the machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory, an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0061] For purposes of providing interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display)) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) by which the user can provide input to the electronic device. Other kinds of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0062] The systems and techniques described herein can be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer having a graphical user interface or a web browser through which the user can interact with an implementation of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0063] A computing system may include a client and a server. The client and the server are generally far from each other and usually interact via a communication network. The relationship between the client and the server is created by computer programs running on respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or a cloud host, which is a host product in the cloud computing service system, solving the defects of difficult management and weak business scalability existing in traditional physical hosts and VPS services.

[0064] In the description of this specification, the description referring to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples", etc. means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0065] Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. A high-precision detection method for track displacement, characterized in that, Including: Obtaining a measurement image and a reference image in real time, where the measurement image is characterized as an image containing measurement targets arranged on an orbit, and the reference image is characterized as an image containing reference targets arranged on a preset reference platform, and the preset reference platform is located on one side of the orbit at the same position as the measurement target; Determining a first displacement amount of the measurement target according to the measurement image obtained at a preset historical moment and a current target moment, and determining a second displacement amount of the reference target according to the reference image obtained at the preset historical moment and the current target moment; Correcting the first displacement amount by the second displacement amount to obtain a target displacement amount of the orbit.

2. The high-precision detection method for track displacement according to claim 1, characterized in that The determining a first displacement amount of the measurement target according to the measurement image obtained at a preset historical moment and a current target moment includes: Based on the initial measurement image obtained at the preset historical moment, determining an initial structural feature image of the measurement target in the initial measurement image, and determining an initial measurement coordinate of a pixel at the center position of the measurement target in the initial structural feature image; Based on the real-time measurement image obtained at the current target moment, determining a real-time structural feature image of the measurement target in the real-time measurement image, and determining a real-time measurement coordinate of a pixel at the center position of the measurement target in the real-time structural feature image; Determining an offset between the initial measurement coordinate and the real-time measurement coordinate, and combining a preset mapping coefficient to determine a first displacement amount of the measurement target at the current target moment, where the mapping coefficient is characterized as a conversion relationship between the pixel range occupied by the imaging of the target object on the camera photosensitive device and its physical size.

3. The high-precision detection method for track displacement according to claim 2, wherein The determining a real-time structural feature image of the measurement target in the real-time measurement image based on the real-time measurement image obtained at the current target moment includes: Obtaining a historical measurement coordinate of a pixel at the center position obtained at the previous moment of the current target moment, and based on the historical measurement coordinate, determining a ROI region of the real-time measurement image, where the ROI region is the region covered by the measurement target; Setting the pixel values of the region outside the ROI region in the real-time measurement image to zero, and performing dynamic threshold binarization on the processed real-time measurement image to obtain a target measurement image; Based on the Hough circle transform, identifying and extracting the structural features of the measurement target in the target measurement image to obtain the real-time structural feature image of the measurement target.

4. The high-precision detection method for track displacement according to claim 2, characterized in that, The determining an offset between the initial measurement coordinate and the real-time measurement coordinate, and combining a preset mapping coefficient to determine a first displacement amount of the measurement target at the current target moment includes: According to the horizontal coordinate offset and the vertical coordinate offset between the initial measurement coordinate and the real-time measurement coordinate, combining a preset mapping coefficient to determine the horizontal displacement and the vertical displacement of the measurement target, and taking the horizontal displacement and the vertical displacement as the first displacement amount of the measurement target at the current target moment.

5. The high-precision detection method for track displacement according to claim 2, characterized in that, The determining steps of the mapping coefficient include: Performing sub-pixel accuracy correction on the pixels at the edge position of the measurement target in the initial structural feature image; Based on the corrected initial structural feature map, determine the pixel radius of the measurement target in the initial structural feature image; Obtain the true radius of the measurement target, and use the ratio of the true radius to the pixel radius as the mapping coefficient.

6. The high-precision detection method for track displacement according to claim 5, characterized in that, Perform sub-pixel accuracy correction on the pixels at the edge position of the measurement target in the initial structural feature image; Based on the corrected initial structural feature image, determining the pixel radius of the measurement target in the initial structural feature image includes: Based on the edge difference method, the gray center of gravity method, and the fitting optimization method, respectively perform sub-pixel accuracy correction on the pixels at the edge position of the measurement target in the initial structural feature image; Based on the gradient consistency algorithm, according to each corrected initial structural feature image, respectively determine the edge smoothness and consistency degree of each initial structural feature image; Use the initial structural feature image with the highest edge smoothness and consistency degree as the target initial structural feature image, and determine the pixel radius of the measurement target according to the target initial structural feature image.

7. A high-precision detection method for track displacement according to claim 2, characterized in that The determining the real-time measurement coordinates of the pixels at the center position of the measurement target in the real-time structural feature image includes: Based on the edge difference method, the gray center of gravity method, and the fitting optimization method, respectively perform sub-pixel accuracy correction on the pixels at the edge position of the measurement target in the real-time structural feature image; Based on the gradient consistency algorithm, according to each corrected real-time structural feature image, respectively determine the edge smoothness and consistency degree of each real-time structural feature image; Use the real-time structural feature image with the highest edge smoothness and consistency degree as the target real-time structural feature image, and determine the real-time measurement coordinates of the pixels at the center position of the measurement target according to the target real-time structural feature image.

8. A high-precision detection system for track displacement, characterized in that, Including: An image acquisition module that real-time acquires a measurement image and a reference image. The measurement image is characterized as an image containing a measurement target configured on an orbit, and the reference image is characterized as an image containing a reference target configured on a preset reference platform. The preset reference platform is located on one side of the orbit at the same position as the measurement target; An image processing module that determines the first displacement amount of the measurement target according to the measurement image obtained at a preset historical moment and the current target moment, and determines the second displacement amount of the reference target according to the reference image obtained at the preset historical moment and the current target moment; A displacement correction module that corrects the first displacement amount by the second displacement amount to obtain the target displacement amount of the orbit.

9. A computer-readable storage medium, characterized in that, Instructions are stored in the computer-readable storage medium, and when the instructions run on the terminal device, the terminal device is caused to execute a high-precision detection method for track displacement as described in any one of claims 1 to 7.

10. An electronic device, comprising a memory, a processor, and a program stored on the memory and running on the processor, characterized in that, When the processor executes the program, it implements a high-precision detection method for track displacement as described in any one of claims 1 to 7.

Images