Steel rail displacement measurement target quality detection and correction method
By detecting and correcting the number and distribution of effective pixel points on the rail displacement measurement target image, the problem of target flaw affecting measurement accuracy is solved, and the accuracy of measurement data and the satisfaction of detection quality is achieved.
Patent Information
- Application Number
- CN202411932712.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-26
- Publication Date
- 2025-04-18
AI Technical Summary
The prior art has failed to effectively detect and handle the possible debris of rail displacement measurement targets after long-term use, affecting the displacement measurement accuracy.
Quality detection and correction are performed by measuring the number and distribution of effective pixel points of the target image. For slightly deficient targets, the accuracy of measurement data is achieved; for severely deficient targets, an abnormal alarm is issued.
Effectively identify and deal with target debris, ensure the accuracy of measurement data and the satisfaction of detection quality, and avoid the reduction of displacement measurement accuracy caused by target quality problems.
Smart Images
Figure CN119850563A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the field of rail displacement measurement and relates to a rail displacement measurement target quality detection and correction algorithm. Background Art
[0002] In the field of image measurement, measuring the displacement value of a target object usually requires setting a target image on the target object and analyzing the target displacement in post-processing. In practical applications, as the target object is exposed for a longer time, it will be affected by environmental factors such as rain, snow, and dust coverage, and the quality of the target image will be flawed, which will ultimately affect the displacement measurement accuracy. The factors affecting target quality in different usage environments are complex and difficult to control, and there is no unified quality inspection standard system that comprehensively covers all types of situations. The existing technology does not take into account the possibility that the target will be contaminated after long-term use, and does not effectively judge and deal with this situation.
[0003] At the same time, as the requirements for the displacement detection accuracy of the target object become more and more stringent, and as the application scenarios of image measurement expand, such as in the fields of high-precision mapping and precision industrial manufacturing, the requirements for the quality indicators such as the size accuracy, pattern clarity, and marking stability of the target are constantly increasing. It is necessary to evaluate and correct the quality of the target to ensure the detection accuracy and meet the detection quality index requirements. Summary of the invention
[0004] Based on the above problems, the present invention proposes a method for detecting and correcting the quality of rail displacement measurement targets. By statistically analyzing the number of effective pixels and the distribution of effective pixels in the measurement target image, the quality of the measurement target is evaluated, thereby achieving the technical effect of target quality detection. Further processing is performed based on the evaluation of the measurement target. For slightly contaminated targets, the measurement data can be unbiased. For heavily contaminated targets, target abnormality alarms can be achieved.
[0005] The rail displacement measurement target quality detection and correction method of the present invention has the following specific steps:
[0006] Step 1: Use a camera to capture an image of a qualified target.
[0007] Step 2: Manually select all target images contained in the image acquired by the camera in step 1, and circle the area with a rectangular frame as the qualified target measurement area.
[0008] Step 3: Obtain the number of valid pixels C in the qualified target measurement area selected in step 2 初 And the distribution of effective pixels D 初 .
[0009] Step 4: Use the camera to capture the image of the target to be inspected, and perform the operation in step 2 to select the measurement area of the target to be inspected in the image of the target to be inspected.
[0010] Step 5: Perform step 3 to measure the number C of effective pixels and the distribution D of effective pixels in the target measurement area to be inspected.
[0011] Step 6: According to the number of effective pixels C and the distribution of effective pixels D in the measurement area of the target to be inspected obtained in step 5, the quality of the target to be inspected is evaluated; specifically: according to the number of effective pixels C and the distribution of effective pixels D of the target to be inspected combined with a given threshold, whether the target to be inspected is qualified is determined as follows:
[0012] If |C 初 -C| / C 初 <0.5%, and |D 初 -D| / D 初 <0.5%, the target to be inspected is judged to be qualified, and step 7 is executed; otherwise, the target to be inspected is unqualified, and step 8 is executed.
[0013] Step 7: Extract the center point of the target image in the target measurement area to be inspected and output the measurement result.
[0014] Step 8: If the quality inspection target is found to have quality problems, image processing and correction are performed on the target image in the measurement area of the quality inspection target.
[0015] Step 9: Evaluate the quality of the corrected target again.
[0016] (1) Repeat the steps in step 3 for the target measurement area trimmed by step 8 to obtain the number of effective pixels C in the corrected target measurement area. 修 And the effective pixel distribution D 修 , then proceed to step 8.
[0017] (2) According to the number of effective pixels C and the distribution of effective pixels D in the measurement area of the corrected quality inspection target combined with a given threshold, determine whether the corrected target is qualified; specifically:
[0018] If |C 初 -C| / C 初 <1%, and |D 初 -D| / D 初 If <1%, it is qualified, and then proceed to step 10, otherwise proceed to step 11.
[0019] Step 10: Extract the center point of the corrected target image and output the measurement result.
[0020] Step 11: Output target quality abnormality alarm message.
[0021] The advantages of the present invention are:
[0022] 1. The rail displacement measurement target quality detection and correction method of the present invention can effectively identify the target contamination and alarm to remind maintenance personnel
[0023] 2. The rail displacement measurement target quality detection and correction method of the present invention can achieve the accuracy of measurement data through image processing for slight target contamination. BRIEF DESCRIPTION OF THE DRAWINGS
[0024] Figure 1 The present invention is an overall flow chart of the target quality detection and correction method for rail displacement measurement.
[0025] Figure 2 Capturing images for the camera of the present invention;
[0026] Figure 3 Select the mode diagram for the rectangular frame;
[0027] Figure 4 It is a diagram showing the distance between the edge pixels and the center of the target image;
[0028] Figure 5 This is a schematic diagram of the target image center pixel in the rectangular area before replacement;
[0029] Figure 6 This is a schematic diagram of the target influencing factor after replacement in the rectangular area;
[0030] Figure 7 Comparison chart of target image before and after correction. DETAILED DESCRIPTION
[0031] The present invention will be further described in detail below in conjunction with the accompanying drawings.
[0032] The rail displacement measurement target quality detection and correction method of the present invention is as follows: Figure 1 As shown, the specific steps include:
[0033] Step 1: By adjusting the angle of the measuring camera, place the qualified target in the field of view of the measuring camera, and control the measuring camera to take pictures. The number of pixels in the image can represent the size of the target, and the distribution of pixels can represent the shape of the pixels, such as Figure 2 shown.
[0034] Step 2: Manually select all target images contained in the image acquired by the camera in step 1, and use a rectangular frame to enclose the area as the qualified target measurement area. This simplifies the calculation when there are multiple targets in the image, and each qualified target measurement area image is processed as an independent image, such as Figure 3 shown.
[0035] Step 3: Obtain the number and distribution of valid pixels in the qualified target measurement area selected in step 2.
[0036] ① Count the pixels with grayscale values greater than 150 in the measurement area of each qualified target, and record them as the initial effective pixel number C of each qualified target. 初 .
[0037] ②Use the Canny operator-based edge detection algorithm to extract the edge pixel coordinate set of the qualified target measurement area, denoted as P 初 ;P 初 The coordinates of each pixel in the n ,Y n ).
[0038] ③ Use the grayscale centroid method to solve the center point of the qualified target measurement area, recorded as (X 中 ,Y 中 ).
[0039] ④ Calculate P in sequence 初 From each pixel to the center point (X 中 ,Y 中 ) and sum them up, which is recorded as the initial effective pixel distribution D 初 ,like Figure 4 shown.
[0040]
[0041] Where n is all the pixels in P, i = 0, 1, ..., n; X0 and Y0 are P 初 The pixel coordinates.
[0042] Step 4: Place the target to be inspected in the field of view of the measuring camera, perform steps 1 and 2, and select the measurement area of the target to be inspected in the image of the target to be inspected captured by the camera.
[0043] Step 5: Use the method in step 3 to obtain the number and distribution of valid pixels in the measurement area of the target to be inspected selected in step 4.
[0044] Step 6: According to the number C of effective pixels and the distribution D of effective pixels in the measurement area of the target to be inspected obtained in step 5, the quality of the target to be inspected is evaluated.
[0045] 1) According to the number of effective pixels C and the effective pixel distribution D of the target to be inspected, combined with a given threshold, determine whether the target to be inspected is qualified, specifically:
[0046] If |C 初 -C| / C 初 <0.5%, and |D 初 -D| / D 初<0.5%, the target to be inspected is considered qualified and step 7 is performed; otherwise, the target to be inspected is unqualified and step 8 is performed. The above 0.5% threshold is mainly due to the reasonable pixel fluctuation that may be caused by the change of light source, which does not affect the center point coordinates.
[0047] Step 7: Extract the center point of the target image in the target measurement area to be inspected and output the measurement result.
[0048] Step 8: If the quality inspection target detects quality problems, the target image in the measurement area of the quality inspection target is processed and corrected. The steps are as follows:
[0049] ①Define a rectangular area of 5*5 pixels.
[0050] ② Slide the rectangular area on the target measurement area to be inspected. For each pixel position on the target measurement area to be inspected, compare the rectangular area with the target measurement area to be inspected pixel by pixel, find the maximum value of the pixel value in the rectangular area, and use the maximum value found as the new pixel value of the pixel at the center point of the rectangular area in the target measurement area to be inspected. Figure 5 The figure shows all 25 pixels of the entire rectangular area at a certain position on the target measurement area to be inspected. The pixel value at the center of the rectangular area in the target measurement area to be inspected is 10, so its pixel value is replaced with the maximum value of 80 among all 25 pixels in the rectangular area. After the replacement, the pixel value in the rectangular area is as follows: Figure 6 shown.
[0051] ③After all the pixel values in the target measurement area to be inspected are replaced in ②, slide the rectangular area on the image again. For each pixel position, compare the rectangular area with the image pixel by pixel to find the minimum pixel value in the rectangular area. The minimum value found is used as the new pixel value of the current pixel position. The final result is as follows: Figure 7 shown.
[0052] Through the above operation, after traversing all pixels in the entire target measurement area to be inspected, the target image diameter in the target measurement area will become larger, but because the target uses its center point to calculate the displacement, the increase in the target diameter will not affect the displacement measurement result.
[0053] Step 9: Evaluate the quality of the corrected target again.
[0054] (1) Repeat the steps in step 3 for the quality inspection target measurement area trimmed by step 8 to obtain the number of effective pixels C in the corrected quality inspection target measurement area. 修 And the effective pixel distribution D 修 , then proceed to step 8.
[0055] (2) According to the number of effective pixels C and the distribution of effective pixels D in the measurement area of the corrected quality inspection target combined with a given threshold, it is determined whether the corrected target is qualified, specifically:
[0056] If |C 初 -C| / C 初 <1%, and |D 初 -D| / D 初 If <1%, it is qualified and the procedure is to proceed to step 10, otherwise, the procedure is to proceed to step 11. The 1% threshold is the maximum allowable deviation within the measurement accuracy requirement range.
[0057] Step 10: Extract the center point of the corrected target image and output the measurement result.
[0058] Step 11: Output target quality abnormality alarm message.
Claims
1. A method for detecting and correcting the quality of a rail displacement measurement target, characterized in that: The steps are: Step 1: Use a camera to capture an image of a qualified target; Step 2: Manually select all target images contained in the image acquired by the camera in step 1, and circle the area with a rectangular frame as the qualified target measurement area; Step 3: Obtain the number of valid pixels C in the qualified target measurement area selected in step 2 初 And the distribution of effective pixels D 初 ; Step 4: Use a camera to capture an image of the target to be inspected, and perform the operation in step 2 to select a measurement area of the target to be inspected in the image of the target to be inspected; Step 5: Perform step 3 to measure the number C of effective pixels and the distribution D of effective pixels in the target measurement area to be inspected; Step 6: According to the number of effective pixels C and the distribution of effective pixels D in the measurement area of the target to be inspected obtained in step 5, the quality of the target to be inspected is evaluated; specifically: according to the number of effective pixels C and the distribution of effective pixels D of the target to be inspected combined with a given threshold, whether the target to be inspected is qualified is determined as follows: If |C 初 -C| / C 初 <0.5%, and |D 初 -D| / D 初 <0.5%, the target to be inspected is judged to be qualified, and step 7 is performed; otherwise, the target to be inspected is unqualified, and step 8 is performed; Step 7: Extract the center point of the target image in the target measurement area to be inspected and output the measurement result; Step 8: If the quality inspection target is found to have quality problems, the target image in the measurement area of the quality inspection target is subjected to image processing correction; Step 9: Evaluate the quality of the corrected target again; (1) Repeat the steps in step 3 for the target measurement area trimmed by step 8 to obtain the number of effective pixels C in the corrected target measurement area. 修 And the effective pixel distribution D 修 , then proceed to step 8; (2) According to the number of effective pixels C and the distribution of effective pixels D in the measurement area of the corrected quality inspection target combined with a given threshold, it is determined whether the corrected target is qualified. The determination method is as follows: If |C 初 -C| / C 初 <1%, and |D 初 -D| / D 初 If <1%, it is qualified, then proceed to step 10, otherwise proceed to step 11; Step 10: Extract the center point of the corrected target image and output the measurement result; Step 11: Output target quality abnormality alarm message.
2. A rail displacement measurement target quality detection and correction method as claimed in claim 1, characterized in that: The specific method in step 3 is: 1) Count the pixels with grayscale values greater than 150 in the measurement area of each qualified target, and record them as the initial effective pixel number C of each qualified target. 初 ; 2) Use the Canny operator-based edge detection algorithm to extract the edge pixel coordinate set of the qualified target measurement area, denoted as P 初 ;P 初 The coordinates of each pixel in the n ,Y n ); 3) Use the grayscale centroid method to solve the center point of the qualified target measurement area, recorded as (X 中 ,Y 中 ); 4) Calculate P in sequence 初 From each pixel to the center point (X 中 ,Y 中 ) and sum them up, which is recorded as the initial effective pixel distribution D 初 ; Where i = 0, 1, ..., n, n is all the pixels in P, X0 and Y0 are P 初 The pixel coordinates.
3. A rail displacement measurement target quality detection and correction method as claimed in claim 1, characterized in that: In step 8, the target image correction method is: ①Define a rectangular area of 5*5 pixels; ② Slide the rectangular area on the target measurement area to be inspected. For each pixel position on the target measurement area to be inspected, compare the rectangular area with the target measurement area to be inspected pixel by pixel, find the maximum value of the pixel value in the rectangular area, and use the found maximum value as the new pixel value of the pixel located at the center point of the rectangular area in the target measurement area to be inspected; ③ After all pixel values in the target measurement area to be inspected are replaced in step ②, the rectangular area is slid on the image again. For each pixel position, the rectangular area is compared with the image pixel by pixel to find the minimum pixel value in the rectangular area. The minimum value found is used as the new pixel value of the current pixel position.