Method, device and equipment for calculating maximum shearing amount of circular support and medium

Abstract

The invention provides a method, device and equipment for calculating the maximum shear amount of a circular support and a medium, and relates to the technical field of bridge engineering detection. The method comprises the following steps: acquiring a first image and a second image; wherein the first image is obtained by shooting the side face of the circular support in the first direction, the second image is obtained by shooting the side face of the circular support in the second direction, and the included angle between the first direction and the second direction is larger than 0 degree and smaller than or equal to 90 degrees; after the first image and the second image are preprocessed, the upper edge and the lower edge of the circular support in the first image and the second image are detected respectively; according to the upper and lower edges of the circular support in the first image and the second image, respectively determining edge shearing displacements corresponding to the first direction and the second direction; according to the included angle between the first direction and the second direction, the edge shearing displacement corresponding to the first direction and the second direction is synthesized, and the maximum shearing displacement of the circular support is obtained. According to the invention, the maximum shearing amount of the circular rubber support can be efficiently and accurately measured in real time.

Description

Technical Field

[0001] This invention relates to the field of bridge engineering testing technology, and in particular to a method, apparatus, equipment and medium for calculating the maximum shear capacity of a circular bearing. Background Technology

[0002] Circular rubber bearings are key components in engineering structures such as bridges and buildings, primarily used to absorb vibrations and transfer loads. These bearings are prone to deformation after prolonged compression, requiring shearing treatment. Since shear deformation directly affects the stability and safety of the engineering structure, accurate measurement of its maximum shear value is crucial.

[0003] Traditional measurement methods usually use contact sensors (such as strain gauges) or manual measuring tools (such as vernier calipers), but these methods have the following problems: (1) low accuracy, contact measurement is affected by the sensor installation position and environment, and it is difficult to accurately reflect the overall deformation of the support; (2) complicated operation, manual measurement requires professional personnel to operate, and cannot adapt to large-scale or complex working conditions; (3) unable to monitor in real time, traditional methods are mostly static measurements, and it is difficult to realize dynamic monitoring of the support in actual use.

[0004] In recent years, visual measurement technology has gradually gained attention due to its advantages such as non-contact operation, high precision, and real-time performance. However, most existing visual measurement methods are limited to single-view photography or simple deformation analysis, making it difficult to comprehensively capture the shear deformation characteristics of the bearing, and they lack adaptability to complex working conditions. Therefore, there is an urgent need for a new method that can efficiently, accurately, and in real-time measure the maximum shear of circular rubber bearings. Summary of the Invention

[0005] This invention provides a method, apparatus, equipment, and medium for calculating the maximum shear capacity of a circular bearing, enabling efficient, accurate, and real-time measurement of the maximum shear capacity of a circular rubber bearing.

[0006] In a first aspect, embodiments of the present invention provide a method for calculating the maximum shear capacity of a circular support, including: Acquire a first image and a second image; wherein the first image is obtained by taking a picture of the side of the circular support from a first direction, and the second image is obtained by taking a picture of the side of the circular support from a second direction, and the angle between the first direction and the second direction is greater than 0 degrees and less than or equal to 90 degrees; After preprocessing the first image and the second image, the upper and lower edges of the circular support in the first image and the second image are detected respectively; Based on the upper and lower edges of the circular support in the first and second images, the edge shear displacements corresponding to the first and second directions are determined respectively. Based on the angle between the first direction and the second direction, the edge shear displacements corresponding to the first direction and the second direction are synthesized to obtain the maximum shear displacement of the circular support.

[0007] In one possible implementation, after obtaining the maximum shear displacement of the circular support, the method further includes: Determine the maximum shear strain based on the maximum shear displacement; The degree of deformation of the circular support is evaluated based on the maximum shear strain.

[0008] In one possible implementation, determining the maximum shear strain based on the maximum shear displacement includes: Detect the height of the circular support in the first image or the second image; The ratio of the maximum shear displacement to the height is determined as the maximum shear strain.

[0009] In one possible implementation, detecting the upper and lower edges of the circular support in the first image and the second image respectively includes: detecting the upper and lower edges of the circular support using the Canny edge detection algorithm.

[0010] In one possible implementation, determining the edge shear displacements corresponding to the first direction and the second direction based on the upper and lower edges of the circular support in the first image and the second image, respectively, includes: By identifying the misaligned portion of the upper or lower edge of the circular support in the first and second images respectively, the edge shear displacement corresponding to the first and second directions is obtained.

[0011] In one possible implementation, the preprocessing of the first image and the second image includes: Identify and segment the circular support region in the target image; wherein the target image is the first image or the second image; The target image is sequentially processed by grayscale conversion, Gaussian filtering for noise reduction, and histogram equalization, and its size is calibrated.

[0012] In one possible implementation, identifying and segmenting the circular support region in the target image includes: The target image is input into a preset detection model, which includes a UniRepLKNet backbone network, a PANet neck embedding hypergraph computation module, and an Inner-CIoU loss function optimization module. The UniRepLKNet backbone network expands the feature channels of the target image through an initial convolutional layer and applies large kernel convolution modules layer by layer to extract the global context features of the circular support, outputting multi-scale feature maps. The PANet neck embedding hypergraph computation module performs bottom-up path aggregation on the multi-scale feature maps, outputting a fused feature map. The Inner-CIoU loss function optimization module uses the Inner-CIoU loss function to optimize the bounding boxes of the fused feature map, outputting multiple detection boxes. Based on the confidence level and category label of each detection box, the target detection box corresponding to the circular support is determined, and the circular support region is cut out based on the target detection box.

[0013] Secondly, embodiments of the present invention provide a device for calculating the maximum shear capacity of a circular support, comprising: An acquisition module is used to acquire a first image and a second image; wherein the first image is obtained by taking a picture of the side of the circular support from a first direction, and the second image is obtained by taking a picture of the side of the circular support from a second direction, and the angle between the first direction and the second direction is greater than 0 degrees and less than or equal to 90 degrees; The detection module is used to preprocess the first image and the second image, and then detect the upper and lower edges of the circular support in the first image and the second image respectively. The determining module is used to determine the edge shear displacement corresponding to the first direction and the second direction, respectively, based on the upper and lower edges of the circular support in the first image and the second image; The synthesis module is used to synthesize the edge shear displacements corresponding to the first direction and the second direction based on the angle between the first direction and the second direction, so as to obtain the maximum shear displacement of the circular support.

[0014] Thirdly, embodiments of the present invention provide an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method described in the first aspect or any possible implementation thereof.

[0015] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described in the first aspect or any possible implementation thereof.

[0016] Fifthly, embodiments of the present invention provide a computer program product, including a computer program that, when executed by a processor, implements the method described in the first aspect or any possible implementation thereof.

[0017] The beneficial effects of the embodiments of the present invention compared with the prior art are as follows: This invention proposes a vision-based non-contact measurement method. This method involves capturing side images of a circular support from two different directions (a first direction and a second direction, with an angle greater than 0 degrees and less than or equal to 90 degrees). After preprocessing the first and second images, the upper and lower edges of the circular support in both images are detected, and the shear amount affecting the support deformation is determined, yielding the edge shear displacements corresponding to the first and second directions. Furthermore, based on the near-incompressible nature of the support material, the maximum shear displacement of the circular support is obtained by synthesizing the edge shear displacements corresponding to the first and second directions. This invention provides complete two-dimensional information about the support deformation through images from two directions, and the combination of image processing technology significantly improves measurement accuracy. This method simplifies the shear calculation process, enhances its engineering applicability, and automates image processing and parameter extraction, reducing manual intervention and improving efficiency. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the method for calculating the maximum shear capacity of a circular support provided in an embodiment of the present invention. Figure 2 This is a schematic diagram of the shear displacement of the edge of a circular support provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the detection model provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the maximum shear capacity calculation device for a circular support provided in an embodiment of the present invention; Figure 5 This is a schematic diagram of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0019] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of the invention. However, those skilled in the art will understand that the invention can be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods are omitted so as not to obscure the description of the invention with unnecessary detail.

[0020] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0021] See Figure 1 The document illustrates a flowchart of the implementation of a method for calculating the maximum shear capacity of a circular support according to an embodiment of the present invention, which is described in detail below: Step S101: Obtain a first image and a second image; wherein the first image is obtained by taking a picture of the side of the circular support from a first direction, and the second image is obtained by taking a picture of the side of the circular support from a second direction, and the angle between the first direction and the second direction is greater than 0 degrees and less than or equal to 90 degrees.

[0022] In this embodiment, two cameras can be placed in two different directions on the support (ideally perpendicular to each other along the X and Y axes; the X and Y axes will be used as examples below). The optical axis should be perpendicular to the side of the support, ensuring that the side of the support and the scale are fully within the field of view. A scale is placed vertically next to the support, parallel to the side of the support and located on the same focal plane, for subsequent image dimensional calibration. The position of the light source is adjusted to ensure uniform illumination on the support surface without obvious shadows or reflections, and then images are captured along the X and Y axes respectively.

[0023] Step S102: After preprocessing the first image and the second image, detect the upper and lower edges of the circular support in the first image and the second image respectively.

[0024] Here, preprocessing may include, but is not limited to: Identify and segment the circular support region in the target image; where the target image is either the first image or the second image. The purpose of this step is to accurately locate the circular support in the image, separate the circular support from the background and other objects, and avoid analyzing irrelevant areas.

[0025] The target image is processed sequentially using grayscale conversion, Gaussian filtering for noise reduction, and histogram equalization. Grayscale conversion transforms a color image into a grayscale image, simplifying the data while preserving key contour information. Gaussian filtering eliminates random noise in the image through a smoothing algorithm, preventing noise from affecting target recognition. Histogram equalization enhances the contrast between the support and the background by adjusting the image's brightness distribution.

[0026] Dimensioning is performed using a reference object of known size (such as the ruler mentioned above) to establish the correspondence between pixels and actual dimensions, which facilitates subsequent calculations of edge shear displacement and support height.

[0027] Step S103: Determine the edge shear displacements corresponding to the first direction and the second direction respectively based on the upper and lower edges of the circular support in the first image and the second image.

[0028] For example, such as Figure 2As shown, when a circular support deforms, its side projection in the first and second images is no longer a standard rectangle, but generally appears as a parallelogram. The edge shear displacement here specifically refers to the relative misalignment of the upper or lower edge of the circular support due to deformation. In actual repair operations, this misaligned portion can be used as a reference to precisely shear and correct the deformed support image, thereby restoring its shape to a circle.

[0029] Step S104: Based on the angle between the first direction and the second direction, the edge shear displacements corresponding to the first direction and the second direction are synthesized to obtain the maximum shear displacement of the circular support.

[0030] In actual shearing operations on circular supports, accurately determining the maximum shear displacement is a core prerequisite for ensuring both shearing effectiveness and safety. Relying solely on image data from a single perspective will result in incomplete shear displacement information due to the limitations of the viewpoint, easily leading to insufficient shearing. Therefore, a multi-view shooting strategy is necessary to obtain comprehensive displacement information: first, images of the support are captured from two different directions; then, image analysis tools are used to extract the corresponding edge shear displacements in the two directions; finally, calculations are performed on the two displacements based on the principle of vector synthesis (for example, when the two directions are perpendicular, the edge shear displacement corresponding to the X-axis is Δx, and the edge shear displacement corresponding to the Y-axis is Δy, then the maximum shear displacement is...). This synthesized maximum shear displacement reflects the overall displacement state of the support in space when the deformation amplitude is large. Using this maximum shear displacement as the shearing basis can not only better remove the deformed and misaligned parts of the support, but also avoid excessive shearing beyond the deformation area. Thus, while achieving support shape repair, it ensures the safety and reliability of the shearing operation and avoids unnecessary damage to the original structure of the support.

[0031] Based on the above, this embodiment proposes a vision-based non-contact measurement method. This method involves capturing side images of a circular support (first image and second image) from two different directions (a first direction and a second direction, with an angle greater than 0 degrees and less than or equal to 90 degrees). After preprocessing the first and second images, the upper and lower edges of the circular support in both images are detected, and the shear amount affecting the support deformation is determined, yielding the edge shear displacements corresponding to the first and second directions. Furthermore, based on the near-incompressible nature of the support material, the maximum shear displacement of the circular support is obtained by synthesizing the edge shear displacements corresponding to the first and second directions. This embodiment provides complete two-dimensional information on support deformation through images from two directions, and the combination of image processing technology significantly improves measurement accuracy. This method simplifies the shear calculation process, enhances its engineering applicability, and automates image processing and parameter extraction, reducing manual intervention and improving efficiency.

[0032] In some embodiments, after obtaining the maximum shear displacement of the circular support, the following operations can be performed based on the maximum shear displacement: determining the maximum shear strain based on the maximum shear displacement; and evaluating the degree of deformation of the circular support based on the maximum shear strain.

[0033] Here, the height of the circular support in the first or second image can be detected, and the ratio of the maximum shear displacement to the support height can be calculated to obtain the maximum shear strain for evaluating the mechanical state of the support. The maximum shear strain is the core parameter for judging the working performance of the circular support. Its value directly reflects the degree of deformation of the support. When the maximum shear strain exceeds the preset safety threshold, it indicates that the support has undergone serious deformation, which may affect the stability of the overall structure, and timely repair or replacement is required.

[0034] See Figure 3 As shown, this embodiment designs an LK-Hyper-Net detection model to more accurately identify and segment circular support regions in target images. The model includes a UniRepLKNet backbone network, a PANet neck embedding hypergraph computation module, and an Inner-CIoU loss function optimization module. The model processing procedure is as follows: (1) The UniRepLKNet backbone network expands the feature channels of the target image through an initial convolutional layer, and applies large kernel convolutional modules layer by layer to extract the global context features of the circular support, and outputs multi-scale feature maps.

[0035] Specifically, color images along the X and Y axes are used as input, and the input image is a 24-bit RGB image (a matrix with width W and height H). If the image size does not match the model's input requirements (e.g., the default 640x640), it can be scaled using bilinear interpolation while maintaining the aspect ratio, and zero padding can be added if necessary to match the square input. The UniRepLKNet backbone first expands the image channels from 3 to an initial 64 feature channels using an initial convolutional layer (kernel size 3x3, stride 1, padding 1). Subsequently, large kernel convolutional modules are applied layer by layer: for each large kernel convolutional block, a kernel size of 13x13 is used to expand the receptive field to capture the global context of the circular support. Multi-scale feature maps are output, with shallow feature maps capturing edge details and deep feature maps capturing semantic information (overall deformable regions). The multi-scale feature maps output by the backbone are then fed into the PANet neck.

[0036] Corresponding to Figure 3The backbone part (UniRepLKNet is a backbone architecture) implements large kernel convolution operations (kernel size 13x13) through the hierarchical structure from Stage 1 to Stage 4, combined with SmaKBlock and LarKBlock. P2, P3, P4, and P5 are multi-scale feature maps output by the backbone network.

[0037] (2) The PANet neck embedding hypergraph computation module performs bottom-up path aggregation on the multi-scale feature maps and outputs the fused feature map.

[0038] This module uses upsampling (2x2 transposed convolution) and lateral connections to fuse high-resolution shallow features with low-resolution deep features. Subsequently, the embedded hypergraph computation module treats the feature map as nodes of a hypergraph, with each node corresponding to a feature vector. Hyperedges are constructed through higher-order correlation calculations (using the K-nearest neighbor algorithm, K=8, to connect nodes). Then, a hypergraph convolution operation is applied, and for each hyperedge, attention weights between nodes are calculated (using softmax normalization). Finally, higher-order features (second- and third-order correlations) are aggregated to refine the multi-scale fusion, outputting the fused feature map. Corresponding to Figure 3 The Neck part (PANet is a commonly used Neck structure) achieves bottom-up path aggregation through the Decoder Pathway and the Bottom-up Augmentation Pathway. Hyper-MAN③ performs Distance-Based Hypergraph Construction → Hypergraph Convolution (HyperConv) → Hybrid Aggregation (MANet①), which is the core logic of the neck hypergraph computation module. C2f② and C32k④ are feature extraction blocks, and Hypergraph is the hypergraph.

[0039] (3) The Inner-CIoU loss function optimization module uses the Inner-CIoU loss function to optimize the bounding boxes of the fused feature map, outputs the position, confidence and category label of multiple detection boxes, determines the target detection box corresponding to the circular support, and cuts out the circular support region based on the target detection box.

[0040] Corresponding to Figure 3 The detection head portion employs an anchor-free design, directly predicting bounding box coordinates, confidence scores, and class labels. It uses the Inner-CIoU loss function for bounding box optimization, calculating the Intersection over Union (IoU) between the predicted and ground truth boxes, and dynamically adjusting the weights of low-IoU samples (through a focus loss mechanism and weighting factors). α =0.25, γ =2.0), to improve regression accuracy for small targets or overlapping regions. Output N3, N4, N5, which are the coordinates (x_min, y_min, x_max, y_max), confidence score (threshold set to 0.5), and class label of each detection box.

[0041] Non-maximum suppression (IoU threshold 0.45) is applied to all candidate bounding boxes output by the detection head to suppress overlapping boxes and retain the highest confidence boxes. For the identified support region, its bounding box coordinates are extracted, and then semantic segmentation is performed. Based on the features within the detection box, an additional segmentation branch (1x1 convolution followed by upsampling to the original image size) is used to generate a pixel-level mask to segment the precise contour of the support. Finally, the extended region of the mask bounding box is calculated (extended by 10% to include the edges), and a smaller support sub-image is cropped using image slicing operations (output in RGB format, smaller than the original image).

[0042] The following embodiments illustrate a detailed implementation flow of the method for calculating the maximum shear capacity of a circular support according to this application, as detailed below: Step 1: Select two cameras (the default output of photos is 24-bit images with 8 bits / channel x 3 colors) and lenses with the same focal length. Prepare a clearly marked ruler, the length of which covers the height of the support. Position the two cameras along the X and Y axes of the support, respectively, with the optical axis perpendicular to the side of the support, ensuring that the side of the support and the ruler are fully in the field of view. Place the ruler vertically (with the scale horizontal) next to the support, parallel to the side of the support and on the same focal plane. Adjust the position of the light source to ensure uniform lighting on the support surface, without obvious shadows or reflections, and then take images along the X and Y axes respectively.

[0043] Step two: The acquired color images along the X and Y axes are used as inputs and fed into the LK-Hyper-Net detection model described in the above embodiment. The pre-trained LK-Hyper-Net model parameters are loaded to identify and segment the circular support regions in the images.

[0044] Step 3, Image Grayscale Conversion: This involves processing each pixel in the color image. This is done using two nested loops: one iterates through the image's rows (y-coordinates), and the other iterates through the images' columns (x-coordinates). For the currently processed pixel, the values ​​of its corresponding R, G, and B channels are obtained. If the current pixel coordinates are (x, y), its R(x,y), G(x,y), and B(x,y) values ​​are obtained, and then the human eye brightness standard weighted formula is applied. Calculate the grayscale value Then, the calculated integer gray value Gray(x,y) is assigned to the pixel at the corresponding coordinates (x, y) in the output grayscale image. This process is repeated for all pixels in the image until all pixels are processed, resulting in the final single-channel grayscale image I. gray .

[0045] Step 4, Gaussian filtering for noise reduction: Apply Gaussian filtering to each single-channel grayscale image with a kernel size of 7×7 and a standard deviation σ = 1.0 to reduce noise interference.

[0046] Gaussian filter formula:

[0047] In the formula, The coordinates are relative to the center of the filter kernel, which is usually defined as (0,0). σ is the standard deviation, which controls the width of the Gaussian distribution. The larger σ is, the wider the filter kernel, the smoother the distribution, and the stronger the smoothing effect, but the more details are lost. The Gaussian function is a normalization constant to ensure that the total weights of the filter kernel sum to 1, avoiding changes to the overall brightness of the image. The Gaussian function determines the weight at each location in the filter kernel, with the center point having the largest weight and the weight gradually decreasing towards the center. This characteristic allows Gaussian filtering to focus more on local pixels during smoothing, preserving the structural information of the image.

[0048] The Gaussian filter kernel is a matrix whose elements are calculated using a Gaussian function and used in subsequent convolution operations. The kernel is an odd-numbered matrix to ensure central symmetry, and its size is related to σ. The formula for the kernel size is: This is to cover the main energy of the Gaussian function (approximately 99.7%). The filter kernel is centered at... Calculate the coordinates of each grid point. Substitute the coordinates into the Gaussian function. We obtain the corresponding weights, sum all the weights to get the total S. We then divide each weight by S to ensure that the sum of the filter kernels is 1.

[0049] For example, for a 3×3 filter kernel, the coordinates are: ,calculate = ≈0.0585. After calculating all weights sequentially, sum all weights to obtain the total S. Divide each weight by S to get: This process is a normalization process. After normalization, the weights sum to 1, which avoids changing the image brightness during convolution.

[0050] Gaussian filtering applies a filter kernel to an image through a convolution operation, calculating a new value for each pixel.

[0051] Convolution formula:

[0052] In the formula, These are the pixel values ​​of the original image; These are the filtered pixel values; For the filter kernel at position The weights; k is the radius of the filter kernel (e.g., k=1 for a 3×3 kernel).

[0053] Place the filter kernel on each pixel of the image, centered at that pixel. Multiply each weight of the filter kernel by the corresponding image pixel value. Summate the products to obtain the new pixel value at that location. Repeat this process for all pixels in the image to obtain the Gaussian-filtered image I. filtered-gray At image edges, the filter kernel may extend beyond the image's boundaries, requiring special handling. Here, we use mirror filling along the edge axis to maintain continuity and reduce edge artifacts.

[0054] For example, for the pixel values ​​of a 3×3 region: Using a normalized 3×3 Gaussian filter kernel (as in the example above), the new value .

[0055] Step 5: Size calibration: On the reference image I filtered-gray Identify the scale markings on the ruler and measure the number of pixels l corresponding to a known length L. Calculate the conversion factor. .

[0056] The preprocessed color image is converted from the RGB color space to the HSV (Hue, Saturation, Value) color space. An HSV threshold range for green is defined. This needs to be adjusted based on the green values ​​on the actual image's scale, for example: setting a Hue range (the H value range corresponding to green), a Saturation range (to avoid colors that are too light or too dark), and a Value range (to avoid colors that are too dark or too bright). Color thresholding is applied: a binary mask is generated based on the defined HSV range. Pixels in the image whose H, S, and V values ​​are all within the defined range are marked as foreground (e.g., white, value = 255) in the mask, and other pixels are marked as background (e.g., black, value = 0). This mask theoretically only includes the main area of ​​the scale. Considering that black typically corresponds to a very low V value in the HSV space, a threshold range for low V values ​​is defined. V value thresholding is applied: the V channel of the HSV image is thresholded to generate a binary mask representing low-brightness (potentially black) regions. Perform a logical AND operation between the scale region mask and the low-brightness region mask. This yields a new binary mask containing only black pixels within the green scale region, corresponding to the tick marks. Use the resulting tick mark candidate pixel mask. Apply a probabilistic Hough transform to detect line segments in the image. Retain straight lines within the scale region, filtering out weaker or shorter segments based on their length, keeping only the sharp, primary tick marks. Sort the filtered horizontal tick marks according to their vertical position, with the length L to be measured corresponding to the first sharp tick mark at the top of the scale and the last sharp tick mark at the bottom. Select the two lines with the smallest and largest y-coordinates after sorting as lines P1 and P2, respectively. Determine the representative coordinate points P1=(x1,y1) and P2=(x2,y2) for the two selected target tick marks. Calculate the pixel distance between the two selected points P1 and P2. The actual values ​​of the two markings identified by the machine are manually read, and the difference L is taken. This difference is then used to obtain the conversion factor. .

[0057] Step 6, Histogram equalization: This process is applied to the denoised image I. filtered-gray Histogram equalization is applied to enhance the contrast between the support edges and the background. Let the total number of pixels in the image be N, and iterate through every pixel in the denoised image. Count the number of pixels at each gray level k (where k ranges from 0 to 255), denoted as n. k Thus, the grayscale histogram H(k) = n of the image is obtained. k The number of pixels n for each grayscale level. k Dividing by the total number of pixels N, we obtain the probability p(k) of each gray level. This yields the probability density function of the image. , where k=0,1,...,255. The cumulative distribution function CDF is the cumulative sum of the probability density function PDF. For gray level k, its cumulative distribution function CDF(k) represents the proportion of pixels with gray values ​​less than or equal to k. The range of CDF(k) is from arrive The core of histogram equalization is to find a transformation function T(k) that maps the original gray level k to a new gray level k′, making the distribution of the new gray level k′ as uniform as possible. This transformation function directly utilizes the CDF calculated in the previous step.

[0058] The mapping formula is: .

[0059] Where L is the total number of gray levels (for an 8-bit grayscale image, L=256, so L-1=255). CDF(k) is the cumulative distribution function value of the original gray level k (range 0 to 1). round() indicates rounding to the nearest integer, because the new gray level must be an integer.

[0060] This transformation function directly maps the cumulative probability of pixels in the original image to a new grayscale. Originally dense grayscale regions (where CDF grows rapidly) are stretched to a wider new grayscale range; originally sparse grayscale regions (where CDF grows slowly) are compressed. The overall effect is to make the histogram of the output image flatter (uniformly distributed).

[0061] Create a new blank image with the same size as the original image. Iterate through each pixel of the original (denoised) image again, reading the original grayscale value k of that pixel. Using the previously obtained transformation function T(k), calculate the new grayscale value k′=T(k). Write this new grayscale value k′ to the corresponding pixel in the new image. After mapping all pixels, you will obtain the histogram-equalized image I. enhanced-gray .

[0062] After histogram equalization, the overall contrast of the image is significantly enhanced. For the support image, the previously blurry support edges (because their gray values ​​are likely similar to the background) become clearer and more prominent because their gray value difference with the background is amplified. This provides a better foundation for subsequent edge detection algorithms to accurately locate the edges.

[0063] Step 7: Apply the Canny edge detection algorithm to the preprocessed image, setting a low threshold of 50 and a high threshold of 150 to extract the edges of the top and bottom of the support.

[0064] The Sobel operator is used to calculate the gradient of the image in the horizontal (X) and vertical (Y) directions.

[0065] The Sobel core is located in the horizontal direction. Vertical direction .

[0066] Calculate the gradient magnitude for each pixel. A larger G value indicates that the point is more likely to be an edge. Calculate the gradient direction for each pixel: , Indicates the direction of the edge. Outputs the gradient magnitude image G and the gradient direction image. Traverse the gradient magnitude image G. For each pixel, examine its gradient direction θ. Compare the gradient magnitude G of that pixel with the gradient magnitudes of its two neighboring pixels along the gradient direction. If the pixel's G value is not a local maximum along its gradient direction, set its magnitude to zero (suppress); otherwise, leave it unchanged. This results in an image B containing thinned edge candidate points (many false and noisy edges have been suppressed). Traverse image B, if the gradient magnitude G of a pixel > T... high If the gradient magnitude T of the pixel is strong, it is marked as a strong edge; low <G≤T high If the gradient magnitude G of a pixel is less than or equal to T, it is marked as a weak edge. low If a pixel is not a strong edge, it is marked as a non-edge (suppressed), and the final output is an image C that marks strong and weak edge pixels. All strong edge pixels are directly determined as final edges. All weak edge pixels are checked; if a weak edge pixel has a strong edge pixel (or a weak edge pixel already determined as a final edge) within its 8-neighborhood, then that weak edge pixel is also determined as a final edge. This process is repeated until no weak edge can be connected to the strong edge chain. All weak edge pixels not connected to the strong edge chain are ultimately suppressed (treated as non-edges). The output is a binary edge image I containing the clear edge contours of the top and bottom of the support. edge .

[0067] Step 8: In the X-axis or Y-axis image, measure the vertical pixel distance h between the top and bottom edges. In the X-axis deformed image, measure the horizontal pixel difference between the top and bottom edges. ,calculate In the Y-axis deformed image, the horizontal pixel difference is measured and calculated. Calculate the shear angles along each axis. , Calculate the total shear displacement. Calculate the total shear angle .

[0068] Step 9: Based on the approximate incompressibility of rubber material, calculate the maximum shear strain using the extracted parameters:

[0069] The maximum shear strain, as a characterization of the maximum shear amount, is used to assess the degree of deformation of the support.

[0070] In a more specific embodiment, two 2-megapixel industrial cameras with a focal length of 25 mm are used, positioned along the X and Y axes of the support, respectively. The support has a diameter of 100 mm, an initial height of 50 mm, and a scale with 1 mm increments. Through image capture and preprocessing, a conversion factor k = 0.1 mm / px is calculated. The Canny algorithm is used to extract edges, resulting in a height h = 50 mm and a shear displacement Δ. x =5mm, Δ y =3mm. Final calculation of total shear displacement. Simultaneously calculate the maximum shear strain. The results showed that the support underwent slight shear deformation.

[0071] Based on the above embodiments, the technical effects achievable by the method of this application include: avoiding the limitations of traditional contact methods through non-contact measurement, making it suitable for real-time monitoring and complex working conditions; the orthogonal vision system provides complete two-dimensional information on support deformation, and the combination with image processing technology significantly improves measurement accuracy; utilizing the incompressible properties of rubber materials simplifies the shear calculation process and enhances the engineering applicability of the method; image processing and parameter extraction can be automated, reducing manual intervention and improving efficiency.

[0072] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0073] The following are embodiments of the apparatus of the present invention. For details not described in detail, please refer to the corresponding method embodiments described above.

[0074] Figure 4 A schematic diagram of the structure of the maximum shear capacity calculation device for a circular support provided in an embodiment of the present invention is shown. For ease of explanation, only the parts related to the embodiment of the present invention are shown. Figure 4 As shown, the maximum shear capacity calculation device 4 for the circular support includes: The acquisition module 41 is used to acquire a first image and a second image; wherein the first image is obtained by taking a picture of the side of the circular support from a first direction, and the second image is obtained by taking a picture of the side of the circular support from a second direction, and the angle between the first direction and the second direction is greater than 0 degrees and less than or equal to 90 degrees. Detection module 42 is used to preprocess the first image and the second image, and then detect the upper and lower edges of the circular support in the first image and the second image respectively. The determining module 43 is used to determine the edge shear displacement corresponding to the first direction and the second direction respectively based on the upper and lower edges of the circular support in the first image and the second image; The synthesis module 44 is used to synthesize the edge shear displacements corresponding to the first direction and the second direction based on the angle between the first direction and the second direction, so as to obtain the maximum shear displacement of the circular support.

[0075] In one possible implementation, after obtaining the maximum shear displacement of the circular support, the synthesis module 44 is further configured to: Determine the maximum shear strain based on the maximum shear displacement; The degree of deformation of the circular support is evaluated based on the maximum shear strain.

[0076] In one possible implementation, the synthesis module 44 is used for: Detect the height of the circular support in the first image or the second image; The ratio of the maximum shear displacement to the height is determined as the maximum shear strain.

[0077] In one possible implementation, the detection module 42 is used to detect the upper and lower edges of the circular support using the Canny edge detection algorithm.

[0078] In one possible implementation, the determining module 43 is used for By identifying the misaligned portion of the upper or lower edge of the circular support in the first and second images respectively, the edge shear displacement corresponding to the first and second directions is obtained.

[0079] In one possible implementation, the detection module 42 is used for: Identify and segment the circular support region in the target image; wherein the target image is the first image or the second image; The target image is sequentially processed by grayscale conversion, Gaussian filtering for noise reduction, and histogram equalization, and its size is calibrated.

[0080] In one possible implementation, the detection module 42 is used for: The target image is input into a preset detection model, which includes a UniRepLKNet backbone network, a PANet neck embedding hypergraph computation module, and an Inner-CIoU loss function optimization module. The UniRepLKNet backbone network expands the feature channels of the target image through an initial convolutional layer and applies large kernel convolution modules layer by layer to extract the global context features of the circular support, outputting multi-scale feature maps. The PANet neck embedding hypergraph computation module performs bottom-up path aggregation on the multi-scale feature maps, outputting a fused feature map. The Inner-CIoU loss function optimization module uses the Inner-CIoU loss function to optimize the bounding boxes of the fused feature map, outputting multiple detection boxes. Based on the confidence level and category label of each detection box, the target detection box corresponding to the circular support is determined, and the circular support region is cut out based on the target detection box.

[0081] Based on the above, this embodiment proposes a vision-based non-contact measurement method. This method involves capturing side images of a circular support (first image and second image) from two different directions (a first direction and a second direction, with an angle greater than 0 degrees and less than or equal to 90 degrees). After preprocessing the first and second images, the upper and lower edges of the circular support in both images are detected, and the shear amount affecting the support deformation is determined, yielding the edge shear displacements corresponding to the first and second directions. Furthermore, based on the near-incompressible nature of the support material, the maximum shear displacement of the circular support is obtained by synthesizing the edge shear displacements corresponding to the first and second directions. This embodiment provides complete two-dimensional information on support deformation through images from two directions, and the combination of image processing technology significantly improves measurement accuracy. This method simplifies the shear calculation process, enhances its engineering applicability, and automates image processing and parameter extraction, reducing manual intervention and improving efficiency.

[0082] Figure 5 This is a schematic diagram of an electronic device provided in an embodiment of the present invention. For example... Figure 5 As shown, the electronic device 5 of this embodiment includes a processor 50 and a memory 51. The memory 51 stores a computer program 52. When the processor 50 executes the computer program 52, it implements the steps in the various method embodiments described above. Alternatively, when the processor 50 executes the computer program 52, it implements the functions of each module in the various device embodiments described above.

[0083] For example, computer program 52 may be divided into one or more modules / units, which are stored in memory 51 and executed by processor 50 to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of computer program 52 in electronic device 5.

[0084] Electronic device 5 may include, but is not limited to, processor 50 and memory 51. Those skilled in the art will understand that... Figure 5 This is merely an example of electronic device 5 and does not constitute a limitation on electronic device 5. It may include more or fewer components than shown, or combine certain components, or different components. For example, electronic device 5 may also include input / output devices, network access devices, buses, etc.

[0085] The processor 50 can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0086] The memory 51 can be an internal storage unit of the electronic device 5, such as a hard disk or RAM. The memory 51 can also be an external storage device of the electronic device 5, such as a plug-in hard disk, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card. Furthermore, the memory 51 can include both internal and external storage units of the electronic device 5. The memory 51 is used to store the computer program 52 and other programs and data required by the electronic device 5. The memory 51 can also be used to temporarily store data that has been output or will be output.

[0087] For the sake of simplicity and clarity, only the above-described functional modules / units are used as examples. In practical applications, the functions described above can be assigned to different functional modules / units as needed. These modules / units can be implemented in hardware, software, or a combination of both.

[0088] This invention also provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the methods described in the above-described method embodiments.

[0089] This invention also provides a computer program product, including a computer program. When the computer program is executed by a processor, it implements the methods described in the above-described method embodiments.

[0090] Computer programs include computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. Computer-readable media can include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0091] In the above embodiments, the descriptions of each embodiment have their own emphasis. Parts not detailed or described in a particular embodiment can be referred to in the relevant descriptions of other embodiments. Unless otherwise specified or in conflict with logic, the terminology and / or descriptions between different embodiments are consistent and can be referenced interchangeably. Technical features in different embodiments can be combined to form new embodiments based on their inherent logical relationships.

[0092] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for calculating the maximum shear capacity of a circular support, characterized in that, include: Acquire a first image and a second image; wherein the first image is obtained by taking a picture of the side of the circular support from a first direction, and the second image is obtained by taking a picture of the side of the circular support from a second direction, and the angle between the first direction and the second direction is greater than 0 degrees and less than or equal to 90 degrees; After preprocessing the first image and the second image, the upper and lower edges of the circular support in the first image and the second image are detected respectively; Based on the upper and lower edges of the circular support in the first and second images, the edge shear displacements corresponding to the first and second directions are determined respectively. Based on the angle between the first direction and the second direction, the edge shear displacements corresponding to the first direction and the second direction are synthesized to obtain the maximum shear displacement of the circular support.

2. The method for calculating the maximum shear capacity of a circular support according to claim 1, characterized in that, After obtaining the maximum shear displacement of the circular support, the process further includes: Determine the maximum shear strain based on the maximum shear displacement; The degree of deformation of the circular support is evaluated based on the maximum shear strain.

3. The method for calculating the maximum shear capacity of a circular support according to claim 2, characterized in that, Determining the maximum shear strain based on the maximum shear displacement includes: Detect the height of the circular support in the first image or the second image; The ratio of the maximum shear displacement to the height is determined as the maximum shear strain.

4. The method for calculating the maximum shear capacity of a circular support according to claim 1, characterized in that, The step of detecting the upper and lower edges of the circular support in the first image and the second image respectively includes: detecting the upper and lower edges of the circular support using the Canny edge detection algorithm.

5. The method for calculating the maximum shear capacity of a circular support according to claim 1, characterized in that, The step of determining the edge shear displacements corresponding to the first direction and the second direction based on the upper and lower edges of the circular support in the first image and the second image, respectively, includes: By identifying the misaligned portion of the upper or lower edge of the circular support in the first and second images respectively, the edge shear displacement corresponding to the first and second directions is obtained.

6. The method for calculating the maximum shear capacity of a circular support according to any one of claims 1 to 5, characterized in that, The preprocessing of the first image and the second image includes: Identify and segment the circular support region in the target image; wherein the target image is the first image or the second image; The target image is sequentially processed by grayscale conversion, Gaussian filtering for noise reduction, and histogram equalization, and its size is calibrated.

7. The method for calculating the maximum shear capacity of a circular support according to claim 6, characterized in that, The process of identifying and segmenting the circular support region in the target image includes: The target image is input into a preset detection model, which includes a UniRepLKNet backbone network, a PANet neck embedding hypergraph computation module, and an Inner-CIoU loss function optimization module. The UniRepLKNet backbone network expands the feature channels of the target image through an initial convolutional layer and applies large kernel convolution modules layer by layer to extract the global context features of the circular support, outputting multi-scale feature maps. The PANet neck embedding hypergraph computation module performs bottom-up path aggregation on the multi-scale feature maps, outputting a fused feature map. The Inner-CIoU loss function optimization module uses the Inner-CIoU loss function to optimize the bounding boxes of the fused feature map, outputting multiple detection boxes. Based on the confidence level and category label of each detection box, the target detection box corresponding to the circular support is determined, and the circular support region is cut out based on the target detection box.

8. A device for calculating the maximum shear capacity of a circular support, characterized in that, include: An acquisition module is used to acquire a first image and a second image; wherein the first image is obtained by taking a picture of the side of the circular support from a first direction, and the second image is obtained by taking a picture of the side of the circular support from a second direction, and the angle between the first direction and the second direction is greater than 0 degrees and less than or equal to 90 degrees; The detection module is used to preprocess the first image and the second image, and then detect the upper and lower edges of the circular support in the first image and the second image respectively. The determining module is used to determine the edge shear displacement corresponding to the first direction and the second direction, respectively, based on the upper and lower edges of the circular support in the first image and the second image; The synthesis module is used to synthesize the edge shear displacements corresponding to the first direction and the second direction based on the angle between the first direction and the second direction, so as to obtain the maximum shear displacement of the circular support.

9. An electronic device, characterized in that, It includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method as described in any one of claims 1 to 7.

Images