A method of detecting distribution of a copper-based self-lubricating bearing surface lubricant and related apparatus

Abstract

The application discloses a copper-based self-lubricating bearing surface lubricant distribution detection method and related equipment, and the method comprises the following steps: acquiring a bearing target surface image under preset environmental conditions, the preset environmental conditions comprising the same illumination condition and a uniform scanning speed, and the target surface being a surface of the bearing requiring to check the lubricant distribution; performing edge detection on the target surface image, and segmenting out a lubricant distribution area and a defect candidate area, the defect candidate area comprising a plurality of preset defect types; based on the information of the defect types, splicing a lubricant distribution sub-area corresponding to a defect displayed across areas to generate a target detection area; detecting texture row contour points of the target detection area, calculating an inclination angle difference of adjacent texture rows, and in response to the inclination angle difference satisfying a preset splicing condition, triggering a splicing operation; segmenting the lubricant area, generating distribution feature data and mapping to a bearing surface three-dimensional coordinate system, and realizing defect positioning.

Description

Technical Field

[0001] This application relates to the field of image processing technology, and in particular to a method and related equipment for detecting the distribution of lubricant on the surface of a copper-based self-lubricating bearing. Background Technology

[0002] Copper-based self-lubricating bearings primarily use copper alloys as the matrix, which possesses excellent thermal and electrical conductivity as well as a certain level of mechanical strength. During manufacturing, solid lubricants (such as graphite and molybdenum disulfide) are incorporated into the copper matrix through methods such as embedding or mixing. During operation, when the bearing and journal move relative to each other, the solid lubricant is gradually released from the matrix, forming a lubricating film on the friction surfaces. This reduces the coefficient of friction, minimizes wear, and achieves self-lubrication. For example, in mechanical equipment operating in high-temperature, high-dust environments where regular lubrication is not possible, copper-based self-lubricating bearings can maintain stable operation through their own lubrication mechanism.

[0003] The uniformity of lubricant distribution directly affects the performance of copper-based self-lubricating bearings. If the distribution is uneven, some areas of the bearing will experience excessive friction due to lack of lubrication during operation, leading to accelerated wear, increased temperature, and reduced product quality.

[0004] Therefore, there is an urgent need for a method and related device for detecting the distribution of lubricant on the surface of copper-based self-lubricating bearings, which can realize the detection of lubricant distribution on the surface of copper-based self-lubricating bearings. Summary of the Invention

[0005] The main technical problem addressed in this application is to provide a method and related equipment for detecting the distribution of lubricant on the surface of copper-based self-lubricating bearings, thereby enabling the detection of lubricant distribution on the surface of copper-based self-lubricating bearings.

[0006] To address the aforementioned technical problems, this application adopts the following technical solution: The method includes: acquiring an image of the bearing target surface under preset environmental conditions, including identical lighting conditions and a uniform scanning speed; the target surface being the surface of the bearing where the lubricant distribution needs to be inspected; performing edge detection on the target surface image to segment the lubricant distribution area and defect candidate areas, the defect candidate areas including several preset defect types; based on the defect type information, stitching together the lubricant distribution sub-regions corresponding to defects displayed across regions to generate a target detection area; detecting the texture row contour points of the target detection area, calculating the tilt angle difference between adjacent texture rows, and triggering a stitching operation in response to the tilt angle difference satisfying preset stitching conditions; segmenting the lubricant area, generating distribution feature data, and mapping it to the three-dimensional coordinate system of the bearing surface to achieve defect localization.

[0007] Among them, the same lighting conditions include a light source system with a cross layout, which is configured to uniformly illuminate the bearing target surface.

[0008] Specifically, based on the information of the defect type, the lubricant distribution sub-regions corresponding to the defects displayed across regions are spliced ​​together to generate the target detection region. This includes: in response to the fact that the distance between defects is less than or equal to a preset distance threshold and the contour overlap area is greater than or equal to a preset overlap ratio, it is determined to be a nested distribution and a splicing operation is performed.

[0009] The process involves detecting the contour points of the texture rows in the target detection area, calculating the tilt angle difference between adjacent texture rows, and triggering a stitching operation when the tilt angle difference meets the preset stitching conditions. This includes calculating the overlap of the texture row regions as the ratio of the intersection area to the minimum region area; and triggering a stitching operation when the overlap exceeds the preset overlap threshold.

[0010] The calculation of the tilt angle difference between adjacent texture rows includes: extracting the coordinates of the edge contour points on both sides of the texture row; and calculating the tilt angle difference between adjacent texture rows based on the coordinates of the edge contour points on both sides.

[0011] The calculation of the ratio of the intersection area to the minimum area of ​​the texture row region is the overlap degree of the texture row region. This includes: extracting the contour points of the texture row region using an edge detection algorithm; constructing corresponding polygonal geometry for each pair of adjacent texture row regions based on the extracted contour points; obtaining the intersection area by calculating the intersection between the two polygons; and calculating the ratio of the intersection area to the smaller value of the area in each pair of adjacent texture regions to obtain the overlap degree.

[0012] The process includes triggering a splicing operation in response to the tilt angle difference meeting the preset splicing conditions. This operation includes: calculating the tilt angle value of each texture row region, calculating its average value as the first angle value; setting a preset threshold for the angle difference of a single texture row region, comparing the tilt angle of each texture row region with the first angle value, and generating an angle difference; and triggering the calculation of the overlap degree when the angle difference of a certain texture row region exceeds the preset threshold, and splicing the texture row regions based on the overlap degree.

[0013] To address the aforementioned technical problems, another technical solution adopted in this application is: providing a copper-based self-lubricating bearing surface lubricant distribution detection device, which includes: an acquisition module, a segmentation module, a first stitching module, a second stitching module, and a positioning module; the acquisition module: acquires an image of the bearing target surface under preset environmental conditions, including the same lighting conditions and a uniform scanning speed, and the target surface is the surface of the bearing where the lubricant distribution needs to be checked; the segmentation module: performs edge detection on the target surface image, segmenting the lubricant distribution area and the defect candidate area, the defect candidate area including several preset defect types; the first stitching module: based on the defect type information, stitches the lubricant distribution sub-regions corresponding to the defects displayed across regions to generate a target detection area; the second stitching module: detects the texture row contour points of the target detection area, calculates the tilt angle difference between adjacent texture rows, and triggers the stitching operation in response to the tilt angle difference satisfying the preset stitching conditions; the positioning module: segments the lubricant area, generates distribution feature data, and maps it to the three-dimensional coordinate system of the bearing surface to achieve defect positioning.

[0014] To solve the above-mentioned technical problems, another technical solution adopted in this application is: to provide a copper-based self-lubricating bearing surface lubricant distribution detection device, the copper-based self-lubricating bearing surface lubricant distribution detection device comprising: a memory and at least one processor, the memory storing instructions; at least one processor calling the instructions in the memory to cause the copper-based self-lubricating bearing surface lubricant distribution detection device to perform the steps of the copper-based self-lubricating bearing surface lubricant distribution detection method as described above.

[0015] To solve the above-mentioned technical problems, another technical solution adopted in this application is: to provide a computer-readable storage medium storing instructions, which, when processed and executed, implement the steps of the method for detecting the distribution of lubricant on the surface of a copper-based self-lubricating bearing as described above.

[0016] Unlike existing technologies, this application offers the following advantages: By acquiring bearing target surface images under preset environmental conditions, the stability and consistency of image acquisition are ensured. Edge detection and segmentation of the target surface image to separate lubricant distribution areas and defect candidate areas enable rapid and accurate identification of potential defect areas, improving detection efficiency. Based on defect type information, relevant sub-regions are stitched together to generate a target detection area. Combining the detection of texture contour points, calculating tilt angle differences, and performing corresponding stitching operations, the problem of missed detections due to defects displaying across regions is effectively avoided, improving the comprehensiveness and accuracy of detection. Finally, segmenting the lubricant area and generating distribution feature data, mapping it to the three-dimensional coordinate system of the bearing surface, enables defect localization. This allows for a clear and intuitive determination of the specific location of defects on the bearing surface, facilitating timely targeted repair or adjustment. This ensures the performance and service life of copper-based self-lubricating bearings, providing an efficient and accurate detection method for related production and application fields, reducing equipment failure risks, and improving production efficiency and product quality. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating one embodiment of the method for detecting the distribution of lubricant on the surface of copper-based self-lubricating bearings according to this application.

[0018] Figure 2 This is a schematic diagram of the structural framework of one embodiment of the copper-based self-lubricating bearing surface lubricant distribution detection device of this application.

[0019] Figure 3 This is a schematic diagram of the structural framework of one embodiment of the copper-based self-lubricating bearing surface lubricant distribution detection device of this application. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] The terms "comprising" and "having," and any variations thereof, used in the embodiments of this invention are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the steps or units listed, but may optionally include other steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices.

[0022] The terms "first" and "second" used herein are for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more features. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0023] To facilitate understanding of this embodiment, a method for detecting the distribution of lubricant on the surface of a copper-based self-lubricating bearing, as disclosed in this embodiment of the invention, will first be described in detail. For example... Figure 1 As shown, Figure 1 This is a flowchart illustrating one embodiment of the method for detecting the distribution of lubricant on the surface of a copper-based self-lubricating bearing according to this application. The method includes the following steps.

[0024] Step S11: Acquire an image of the bearing target surface under preset environmental conditions. The preset environmental conditions include the same lighting conditions and a uniform scanning speed. The target surface is the surface of the bearing where the distribution of lubricant needs to be checked.

[0025] Specifically, the same lighting conditions include a cross-laid light source system, which can be achieved by setting uniformly distributed LED light sources within a closed detection space. The spectral range of the light sources can include the visible light band of 400-700nm, and the light intensity can be controlled within a reasonable range to reduce shadows and reflections, thereby minimizing their impact on image quality. A uniform scanning speed can be achieved by driving a linear array camera with a high-precision stepper motor. The scanning speed can be set to 5-20 mm / s based on the curvature of the bearing surface and the required detection accuracy, ensuring an overlap rate between 10% and 30% between adjacent scanning lines.

[0026] In some specific embodiments, the device for acquiring images of the bearing target surface may include a linear CCD camera or an area-array CMOS camera. The camera resolution can be set between 1024×768 and 4096×3072 pixels, depending on the size of the bearing surface and the required detection accuracy. During image acquisition, multiple exposure fusion technology can be used to capture the same area 2-5 times with different exposure times. Then, an image fusion algorithm is used to synthesize a high dynamic range image to improve the contrast between the lubricant area and the metal substrate area in the image.

[0027] In some applications, when inspecting small bearings, such as those with a diameter less than 50 mm, the bearing can be fixed on a rotary table. The entire circumferential surface can be imaged through a combination of linear scanning by the camera and the rotation of the table. When inspecting large bearings, such as those with a diameter greater than 200 mm, a multi-camera array can be used to simultaneously scan the bearing surface in segments. Then, image stitching technology can be used to synthesize a complete surface image from these segments. Furthermore, before acquiring the image, the bearing surface can be cleaned, for example, by compressed air blowing or ultrasonic cleaning, to remove dust, oil, and other impurities, improving image clarity and inspection accuracy.

[0028] In some embodiments, the same illumination conditions include a light source system with a cross-layout configuration, the light source system being configured to uniformly illuminate the bearing target surface.

[0029] Specifically, the cross-layout light source system can include two or more sets of linear light sources, each set intersecting at a certain angle, such as between 45° and 135°. Each set of light sources can include multiple LED beads or strip light sources. The luminous intensity of the light source can be precisely controlled by a constant current drive circuit to ensure a uniform illumination intensity distribution on the bearing target surface. The uniformity of the illumination intensity can be controlled within ±5%. The spectral characteristics of the light source system can be selected based on the optical properties of the bearing surface material and the lubricant. For example, for a copper-based bearing surface, a light source with a peak wavelength in the range of 550-650nm can be selected to improve the contrast between the lubricant area and the metal substrate area.

[0030] In some specific embodiments, the light source system can adopt a ring-shaped cross layout, arranging multiple sets of strip light sources in a ring around the bearing target surface. The included angle between adjacent light source groups can be set to 30° to 60°. The illumination direction of each light source group can be at a certain angle to the normal direction of the bearing surface, such as 15° to 45°, to create a diffuse reflection illumination effect and reduce the impact of surface reflection on image acquisition. The driving circuit of the light source can be configured in PWM dimming mode, controlling the luminous intensity of the light source by adjusting the pulse width, achieving intensity regulation within the range of 0 to 100%, with an adjustment accuracy of ±1%.

[0031] In some applications, for bearings with high surface roughness, the illumination angle range of the light source can be increased. For example, a multi-angle composite illumination method can be used to illuminate the bearing surface from multiple directions simultaneously, thereby improving image contrast in areas with surface defects and uneven lubricant distribution. For bearings with complex geometries, a diffuser or softbox can be added to the light source system to scatter the light before it reaches the bearing surface, reducing the effects of shadows and halos. Furthermore, the light source system can be equipped with a light intensity sensor and feedback control circuitry to monitor and adjust the light intensity in real time, ensuring the stability of illumination conditions during long-term testing, with light intensity fluctuations controlled within ±3%.

[0032] Step S12: Perform edge detection on the target surface image to segment the lubricant distribution area and the defect candidate area. The defect candidate area includes several preset defect types.

[0033] Specifically, edge detection can be implemented using classic image processing algorithms, such as the Canny algorithm, the Sobel operator, or the Prewitt operator. Taking the Canny algorithm as an example, its processing includes Gaussian filtering the target surface image to suppress noise, calculating the magnitude and direction of the image gradient, refining edges through non-maximum suppression, and finally determining the final edge using the hysteresis thresholding method. Edge detection can effectively identify the boundary between the lubricant and the metal substrate on the bearing surface, as well as the contour features of potential defects. The image segmentation process can combine thresholding or region growing algorithms. Thresholding segmentation uses the Otsu algorithm to automatically determine the grayscale threshold, dividing the image into lubricant distribution areas and metal substrate areas, while eliminating isolated noise points through morphological operations. Defect candidate regions are extracted based on closed edge regions or grayscale abnormal regions. For example, lubricant missing regions appear as low-grayscale regions with irregular edges, while impurity mixed regions present closed contours with abnormal grayscale value fluctuations. Preset defect types include common defect morphologies such as uneven lubricant distribution, local lubricant missingness, impurities in lubricant regions, surface microcracks, and pits. A feature library is pre-established for each defect type, such as the edge roughness of the lubricant missing area, the gray-level variance of the impurity area, and the continuity of the crack direction, which serve as the basis for subsequent defect identification.

[0034] In some specific embodiments, edge detection algorithms can be optimized by combining them with deep learning models. For example, U-Net or Mask R-CNN networks can be used for semantic segmentation of images to automatically identify lubricant distribution areas and defect candidate regions. Deep learning models are trained on a large number of labeled samples to learn the texture, shape, and grayscale features of different defect types, improving segmentation accuracy and robustness. Image preprocessing steps include histogram equalization or homomorphic filtering to enhance image contrast and improve edge detection performance.

[0035] In some applications, when bearing surfaces have complex textures or reflective areas, a multi-scale edge detection method is employed. This method combines Canny operators with different Gaussian kernel sizes to obtain multi-resolution edge information, and then edge fusion technology is used to generate a complete edge contour. For high-speed online inspection scenarios, the computational efficiency of the edge detection algorithm is optimized. For example, an FPGA-based parallel computing architecture is used to ensure real-time image segmentation and defect candidate region extraction during uniform scanning. For laboratory inspection scenarios, a more refined defect feature database is established by combining manual annotation and algorithm analysis, providing richer training samples for subsequent defect classification.

[0036] Step S13: Based on the defect type information, stitch together the lubricant distribution sub-regions corresponding to the defects displayed across regions to generate the target detection area.

[0037] Specifically, defect type information can include the defect's geometric features, grayscale features, and contour features. For example, uneven lubricant distribution defects may appear as multiple scattered low-grayscale areas, while local lubricant deficiency defects often present as closed contours with irregular edges. Cross-regional defects refer to the same defect being segmented into multiple sub-regions during image segmentation due to noise interference or blurred edges, or sub-regions of adjacent defects being spatially related. The stitching process can begin by extracting the contour coordinates of each sub-region using a contour detection algorithm, and calculating the distance between different sub-regions and the contour overlap area. The sub-region distance can be defined as the shortest distance between the smallest bounding rectangles of two regions, and the contour overlap area can be calculated by measuring the area of ​​the intersection of two contour polygons. When the preset distance threshold and overlap ratio conditions corresponding to the defect type are met, a sub-region stitching operation is triggered, for example, merging the contours of adjacent sub-regions to generate a target detection region containing the complete defect shape.

[0038] In some specific embodiments, the preset defect types can correspond to different splicing strategies. For example, for defects where impurities are mixed in the lubricant region, the corresponding sub-regions may have similar grayscale values ​​and texture features, and the splicing conditions can be set to a lenient spacing threshold and a high overlap ratio. For surface microcrack defects, since the crack direction is continuous, the splicing conditions can be determined by combining the angle difference of the contour direction. When the angle difference of the contour tangents of adjacent sub-regions is less than a preset angle and the spacing is less than the threshold, splicing is performed. In addition, the splicing process can use morphological operations to smooth the edges of the merged regions, eliminating the jagged noise of the contour caused by segmentation errors.

[0039] In some applications, when inspecting high-resolution images or the surface of large bearings, cross-regional defects may be distributed across different image sub-blocks. In such cases, image registration techniques can be used to calibrate the positions of each sub-block before stitching together the cross-block regions based on defect type information. For online real-time inspection scenarios, the stitching algorithm can use heuristic rules to quickly determine the correlation between sub-regions, ensuring timely generation of target detection areas during high-speed scanning. For offline inspection scenarios, manually annotated defect samples can be combined to optimize preset thresholds, improving the stitching accuracy of complex defect morphologies. When the bearing surface has curvature variations, geometric transformation compensation can be introduced during the stitching process. For example, coordinate mapping correction can be performed on the image sub-regions after the surface is unfolded to ensure accurate stitching of cross-regional defects on the curved surface.

[0040] In some embodiments, based on the information of the defect type, the lubricant distribution sub-regions corresponding to the defects displayed across regions are spliced ​​together to generate a target detection region, including: in response to the fact that the distance between defects is less than or equal to a preset distance threshold and the contour overlap area is greater than or equal to a preset overlap ratio, it is determined to be a nested distribution and a splicing operation is performed.

[0041] Specifically, the spacing between defects is determined by calculating the shortest distance between the smallest bounding rectangles of two defect sub-regions. This distance reflects the positional correlation of the sub-regions in image space. The contour overlap area is obtained by extracting the closed contour polygons of the sub-regions and calculating the number of pixels or geometric area of ​​the polygon intersection area, reflecting the degree of overlap in the shape of the sub-regions. The preset spacing threshold and preset overlap ratio are set according to the defect type and detection accuracy. For example, for local lubricant missing defects, the preset spacing threshold can be set to 10-20 pixels, and the preset overlap ratio can be set to 25%-35%. For large areas of unevenly distributed lubricant, the preset spacing threshold can be increased to 30-50 pixels, and the preset overlap ratio can be adjusted to 15%-25%. When the spacing and overlap area of ​​two sub-regions simultaneously meet the preset conditions of the corresponding defect type, they are determined to be nested distributions, triggering a contour merging operation to connect or merge the edge contours of adjacent sub-regions to generate a target detection area containing the complete defect shape.

[0042] In some specific embodiments, the preset threshold can be determined through statistical analysis of defect sample training data. For example, the probability distribution of the sub-region spacing and overlapping area of ​​similar defects in historical detections can be fitted, and a reasonable threshold range can be determined by adding or subtracting the standard deviation from the mean. The calculation of the contour overlap area can be combined with image binarization processing. After converting the sub-regions into binary images, the intersection is obtained through logical AND operation, and then the intersection contour is extracted and the area is calculated using a contour tracking algorithm. When performing the stitching operation, the merged contours can be morphologically smoothed first, such as by using median filtering or Gaussian filtering with a 3×3 template, to eliminate contour burrs or depressions caused by segmentation errors and ensure the continuity and regularity of the target detection area edges.

[0043] In some applications, when inspecting high-resolution images or the surfaces of large bearings, defect stitching across sub-blocks requires first unifying the coordinate reference through image registration techniques. For example, a Scale-Invariant Feature Transform (SIFT) feature point matching algorithm can be used to align adjacent image sub-blocks before calculating the spacing and overlap area. For online real-time inspection scenarios, a preset threshold can be set to a fixed empirical value and embedded with hardware acceleration circuitry to ensure rapid identification of nested distributions during high-speed bearing operation. For offline inspection scenarios, the threshold can be dynamically adjusted based on manually labeled defect samples. For example, the preset overlap ratio for defects containing impurities in the lubricant region can be increased to 35% to avoid mis-stitching due to noise interference. When the bearing surface has curvature, defect spacing calculation requires compensation for surface deformation through three-dimensional coordinate mapping. For example, the cylindrical surface can be unfolded into a planar image before distance measurement to ensure the stitching accuracy of sub-regions with different curvatures.

[0044] Step S14: Detect the contour points of the texture rows in the target detection area, calculate the tilt angle difference between adjacent texture rows, and trigger the stitching operation in response to the tilt angle difference meeting the preset stitching conditions.

[0045] Specifically, the spacing between defects can be determined by calculating the shortest distance between the smallest bounding rectangles of the two defect regions, and the overlap area can be calculated by extracting the contour polygons of each defect region and calculating their intersection area. The preset spacing threshold and preset overlap ratio can be set according to the defect type and detection accuracy requirements. For example, for small local lubricant loss defects, the preset spacing threshold can be set to 10-20 pixels, and the preset overlap ratio to 20%-30%; for larger areas with uneven lubricant distribution, the preset spacing threshold can be appropriately increased to 30-50 pixels, and the preset overlap ratio adjusted to 15%-25%. When the spacing and overlap area of ​​two defect sub-regions simultaneously meet the above conditions, they can be determined to be nested distributions. At this time, the contours of the two sub-regions are merged to generate a target detection area containing the complete defect shape.

[0046] In some specific embodiments, the preset spacing threshold and preset overlap ratio can be adaptively adjusted using machine learning algorithms. For example, by training a random forest model or support vector machine, inputting parameters such as defect type, image resolution, and bearing surface material, the corresponding optimal threshold combination can be output. Furthermore, the calculation of the contour overlap area can be combined with image binarization processing. After converting the defect region into a binary image, the intersection region is obtained through logical AND operation, and then the number of pixels in the intersection region is counted as the overlap area. When performing the stitching operation, the merged contours can first be smoothed, for example, by using Gaussian filtering or Bézier curve fitting to eliminate irregular edges caused by segmentation errors.

[0047] In some applications, when inspecting bearings with curved surfaces, the spacing calculation of defect sub-regions needs to consider the geometric characteristics of the surface. For example, the surface can be unfolded into a planar image through 3D coordinate mapping before distance calculation. For high-speed inspection scenarios, the preset spacing threshold and overlap ratio can be set to fixed empirical values ​​to ensure the real-time performance of the stitching algorithm. For offline inspection scenarios with high accuracy requirements, the thresholds can be finely adjusted for different defect types by manually annotating sample data. For example, the preset overlap ratio for defects containing impurities in the lubricant region can be increased to 35% to avoid incorrect stitching. In addition, when there are multiple nested defects in the image, the stitching operation can be processed sequentially from left to right and from top to bottom to ensure that each target detection area contains only the most correlated sub-region.

[0048] In some embodiments, the texture line contour points of the target detection area are detected, the tilt angle difference between adjacent texture lines is calculated, and a stitching operation is triggered in response to the tilt angle difference meeting a preset stitching condition, including: calculating the ratio of the intersection area to the minimum area of ​​the texture line region as the overlap degree; and triggering the stitching operation in response to the overlap degree exceeding a preset overlap threshold.

[0049] Specifically, the detection of texture row contour points in the target detection region can be achieved through edge detection algorithms, such as using the Canny or Sobel operators to extract the edge contours of the texture rows and obtain the coordinate sequence of the contour points. The tilt angle difference between adjacent texture rows can be calculated by fitting the contour points with a straight line. For example, linear regression can be performed on the left and right edge contour points of each texture row to obtain the slope of the edge line, which is then converted into the tilt angle with respect to the horizontal axis. The angle difference between two adjacent rows is the tilt angle difference. The intersection area can be calculated by converting the texture row region into a binary image, obtaining the overlapping part of adjacent regions through a logical AND operation, and counting the number of pixels in the overlapping area as the intersection area. The minimum area is taken as the smaller value between the areas of two adjacent texture row regions, and the ratio of the two is the overlap degree. When the overlap degree exceeds a preset overlap threshold, it indicates that the adjacent texture row regions have a high correlation, triggering a stitching operation to merge the two.

[0050] In some specific embodiments, the detection of texture row contour points can be optimized by combining morphological operations. For example, the target detection region can be first dilated to connect neighboring pixels, and then continuous contour points can be extracted using a contour tracking algorithm. The calculation of the tilt angle difference can incorporate robust statistical methods, such as the random sampling consensus algorithm, to eliminate interference from abnormal contour points on straight line fitting and improve the accuracy of angle calculation. The preset overlap threshold can be dynamically adjusted based on the width of the texture row and the detection accuracy. For example, for narrow texture rows, the preset overlap threshold can be set to 40%-50%, while for wide texture rows it can be adjusted to 30%-40%.

[0051] In some applications, when detecting bearings with complex curved textures, the calculation of the tilt angle difference of texture rows needs to consider the influence of surface curvature. For example, the surface texture can be mapped to a planar image through 3D coordinate transformation before angle calculation. For high-speed detection scenarios, a GPU-based parallel computing architecture can be used to accelerate contour point detection and overlap calculation, ensuring that the stitching operation is completed in real time during the high-speed rotation of the bearing. To address the issue of low texture contrast due to uneven lubricant distribution, the preprocessing step can include histogram equalization or adaptive threshold segmentation to enhance the edge clarity of texture rows and avoid overlap calculation errors caused by contour detection errors. Furthermore, when multiple texture rows exist in the target detection area, the stitching operation can process them sequentially according to their arrangement, prioritizing the merging of adjacent areas with high overlap and small tilt angle differences to ensure the generation of a continuous and complete lubricant distribution detection area.

[0052] In some embodiments, calculating the tilt angle difference between adjacent texture rows includes: extracting the coordinates of the edge contour points on both sides of the texture row; and calculating the tilt angle difference between adjacent texture rows based on the coordinates of the edge contour points on both sides.

[0053] Specifically, the coordinates of the edge contour points on both sides of a texture row can be extracted using edge detection algorithms, such as the Canny or Sobel operators, to extract the edges of the target detection region and obtain the edge contour point sequences on the left and right sides of the texture row. For each texture row, the left edge contour points can be defined as the set of pixels on the left boundary of the texture row, and the right edge contour points are the set of pixels on the right boundary. When calculating the tilt angle difference based on the coordinates of the two edge contour points, a straight line fitting can be performed on each edge contour point. For example, the least squares method can be used to fit the left and right edge straight lines, calculate the angle between the two lines and the horizontal axis, and take the average of the two angles as the tilt angle of the texture row. The difference in tilt angle between adjacent texture rows is the tilt angle difference.

[0054] In some specific embodiments, the extraction of edge contour points can be optimized by combining morphological filtering, such as removing isolated noise points through erosion and dilation operations to ensure the continuity and integrity of the edge contour. Robust fitting algorithms, such as the random sampling consensus algorithm, can be introduced during the straight-line fitting process to eliminate the interference of abnormal contour points on the fitting results and improve the accuracy of the tilt angle calculation. For texture rows with curved or irregular edges, the edges can be divided into multiple line segments for segmental fitting, and the average value of the tilt angle of each segment can be taken as the overall tilt angle of the texture row.

[0055] In some applications, when inspecting curved bearing surfaces, the coordinates of the edge contour points on both sides of the texture row need to be converted to planar coordinates via 3D coordinate mapping before straight line fitting and angle calculation are performed to compensate for the influence of surface curvature on the tilt angle. For images acquired through high-speed scanning, edge contour point extraction and tilt angle calculation can employ a parallel computing architecture, such as FPGA or GPU-based implementation, to ensure real-time angle difference calculation during the bearing's high-speed rotation. When the texture row edges have low contrast due to uneven lubricant distribution, the preprocessing step can incorporate adaptive threshold segmentation or edge enhancement algorithms to improve the accuracy of contour point extraction and avoid tilt angle calculation errors caused by edge blurring.

[0056] In some embodiments, calculating the ratio of the intersection area to the minimum region area as the overlap degree of the texture row region includes: extracting the contour points of the texture row region using an edge detection algorithm; for each pair of adjacent texture row regions, constructing corresponding polygonal geometry based on the extracted contour points, obtaining the intersection area by calculating the intersection between the two polygons; and performing a ratio operation between the intersection area and the smaller value of the area in each pair of adjacent texture regions to obtain the overlap degree.

[0057] Specifically, when extracting contour points of texture row regions using edge detection algorithms, the Canny algorithm, Sobel operator, or contour tracing algorithm can be used to scan pixel by pixel along the texture row edge to obtain a continuous and closed sequence of contour points. When constructing polygonal geometry, contour points are connected in a clockwise or counterclockwise order to form closed polygons, ensuring the ordered nature of the polygon vertex coordinates for subsequent geometric operations. Calculating the intersection area between two polygons can be achieved using geometric algorithms, such as using a scanline algorithm to calculate the number of pixels in the overlapping region, or using vector operations to solve for the intersection points of polygon edges, determining the vertex coordinates of the intersection region, and then calculating the area. The minimum region area is taken as the smaller value between the areas of two adjacent texture row regions, calculated through pixel statistics or geometric formulas; the ratio of the two is the degree of overlap.

[0058] In some specific embodiments, morphological preprocessing can be performed on the textured region before edge detection. For example, dilation can be used to connect adjacent contour points, and erosion can be used to remove isolated noise points, ensuring that the extracted contour points are complete and continuous. The polygon geometry can be stored as a vertex list, with each vertex containing x and y coordinate information, facilitating subsequent intersection calculation. The intersection area can be calculated by combining image binarization processing. After converting the textured region into a binary image, a logical AND operation is used to obtain the overlapping portion, and then the number of pixels in the overlapping area is counted as the intersection area. Furthermore, for contours with gaps or breaks, interpolation algorithms can be used to supplement missing contour points, ensuring the closure of the polygon.

[0059] In some applications, when detecting subtle textures in high-resolution images, edge detection algorithms can employ sub-pixel precision contour extraction methods to improve the accuracy of contour point coordinates, thereby enhancing the accuracy of overlap calculation. For textures on curved bearing surfaces, the 3D surface coordinates need to be mapped to 2D planar coordinates before constructing polygons to compensate for the influence of surface curvature on contour morphology. In real-time detection scenarios, overlap calculation can utilize hardware acceleration techniques, such as GPU-based parallel polygon operations, to ensure real-time calculation of intersection area and minimum region area during high-speed scanning. For situations where the contrast between the lubricant distribution area and the metal substrate is low, the preprocessing step can incorporate adaptive histogram equalization to enhance the clarity of texture edge lines and avoid overlap calculation errors caused by incomplete contour extraction.

[0060] Step S15: Segment the lubricant area, generate distribution feature data and map it to the three-dimensional coordinate system of the bearing surface to achieve defect location.

[0061] Specifically, segmenting the lubricant region can be achieved using image segmentation algorithms, such as threshold-based segmentation, region growing-based segmentation, or deep learning-based semantic segmentation. By analyzing the grayscale differences, color features, or texture features between the lubricant region and the metal substrate region, the lubricant region is separated from the target surface image. Generating distribution feature data can include calculating geometric features of the lubricant region, such as area, perimeter, centroid position, and shape factor, as well as statistical features such as grayscale mean, variance, and contrast. When mapping these feature data to the bearing surface's three-dimensional coordinate system, it is necessary to first establish a mapping relationship between the image coordinate system and the three-dimensional coordinate system. For example, by obtaining the camera's intrinsic and extrinsic parameters through camera calibration, and combining this with the bearing's geometric model and pose information, the coordinates of the lubricant region in the two-dimensional image are converted into coordinate points in the three-dimensional coordinate system, thereby achieving three-dimensional localization of the defect.

[0062] In some specific embodiments, the image segmentation algorithm can employ the Otsu adaptive thresholding method, automatically determining the segmentation threshold based on the image's grayscale histogram to divide the image into lubricant regions and metal substrate regions. For images with complex textures or uneven illumination, preprocessing techniques such as Gaussian filtering, histogram equalization, or homomorphic filtering can be performed to improve segmentation results. When generating distribution feature data, morphological operations such as dilation, erosion, opening, and closing operations can be used to post-process the segmented regions, removing noise and small holes, and extracting continuous and complete lubricant region contours. When mapping to a three-dimensional coordinate system, a feature point matching-based method can be used. For example, markers with known locations can be set on the bearing surface. By identifying the positions of these markers in the image and combining them with camera calibration parameters, the transformation matrix between the image coordinate system and the three-dimensional coordinate system can be calculated to achieve coordinate mapping.

[0063] The camera calibration process can employ the Zhang Zhengyou calibration method, which involves capturing multiple chessboard images in different poses and calculating the camera's intrinsic parameter matrix, distortion coefficients, and extrinsic parameter matrix. Geometric feature calculations can utilize functions provided by computer vision libraries, such as the `contourArea` function for calculating the contour area and the `moments` function for calculating the centroid. Statistical feature calculations can iterate through all pixels within the lubricant region, statistically analyzing the distribution of grayscale values. Coordinate mapping can be achieved through matrix operations, multiplying image coordinate points by a transformation matrix to obtain the corresponding three-dimensional coordinate points.

[0064] In some applications, when inspecting bearings with complex curved surfaces, it's necessary to consider the impact of the surface's geometric properties on coordinate mapping. For example, a non-uniform rational B-spline (NURBS) surface model can be used to model the bearing surface, accurately mapping the two-dimensional image coordinates onto the three-dimensional surface. For high-speed rotating bearings, to achieve real-time inspection, a linear scan camera combined with synchronous triggering technology can be used to continuously acquire images during bearing rotation. A time synchronization mechanism can then correlate the defect location in the image with the bearing's rotation angle, achieving dynamic defect localization. In industrial environments, to improve system robustness, multi-sensor fusion technology can be employed. For instance, combining laser displacement sensors to acquire three-dimensional point cloud data of the bearing surface with visual inspection results further enhances the accuracy of defect localization.

[0065] In some embodiments, in response to the tilt angle difference meeting a preset splicing condition, a splicing operation is triggered, including: calculating the tilt angle value of each texture row region, calculating its average value as a first angle value; setting a preset threshold for the angle difference of a single texture row region, comparing the tilt angle of each texture row region with the first angle value, and generating an angle difference; in response to the angle difference of a certain texture row region exceeding the preset threshold, triggering the calculation of overlap, and splicing the texture row regions based on the overlap.

[0066] Specifically, the tilt angle of each texture row region can be calculated by fitting a straight line to the edge contour points of the texture row. For example, the least squares method can be used to perform linear regression on the edge contour points on both sides of the texture row to obtain the slope of the edge line, which is then converted into the tilt angle with respect to the horizontal axis. The first angle value is the arithmetic mean of the tilt angles of all texture row regions, reflecting the overall trend of the texture rows within the target detection area. The preset threshold can be set according to the defect type and detection accuracy requirements. For example, for regularly arranged texture rows, the preset threshold can be set to 10°-15°, and for complex curved surfaces or irregular texture rows, the preset threshold can be adjusted to 20°-25°. When the difference between the tilt angle of a texture row region and the first angle value exceeds the preset threshold, it indicates that the direction of the texture row is significantly different from the overall trend. At this time, the overlap calculation is triggered, and the overlap of adjacent texture row regions is used to determine whether to perform a splicing operation.

[0067] In some specific embodiments, the calculation of the tilt angle can be combined with a robust fitting algorithm to eliminate the interference of abnormal contour points on the straight line fitting, thereby improving the accuracy of the angle calculation. The preset threshold can be determined through statistical analysis of historical detection data. For example, calculating the angle difference distribution of a large number of normal texture rows and using the mean plus three times the standard deviation as the preset threshold ensures the reasonableness of the threshold setting. When triggering the overlap calculation, the correlation analysis of texture row regions with angle differences exceeding the threshold and adjacent regions can be prioritized. For example, only the overlap between texture rows with excessive angle differences and their upper and lower adjacent rows can be calculated, reducing unnecessary computation. Furthermore, the stitching operation can be combined with texture consistency detection. For example, when the overlap exceeds the preset threshold and the texture grayscale feature similarity is higher than 80%, the final region merging is performed, avoiding incorrect stitching caused by abnormal angles.

[0068] In the above scheme, by acquiring images of the bearing target surface under preset environmental conditions, the stability and consistency of image acquisition are ensured. By performing edge detection on the target surface image and segmenting the lubricant distribution area and defect candidate area, potential defect areas can be quickly and accurately identified, improving detection efficiency. Based on defect type information, relevant sub-regions are stitched together to generate the target detection area. Combining the detection of the texture line contour points of the target detection area, calculating the tilt angle difference, and performing corresponding stitching operations, the detection omission problem caused by defects displaying across regions is effectively avoided, improving the comprehensiveness and accuracy of detection. Finally, the lubricant area is segmented and the distribution feature data is generated and mapped to the three-dimensional coordinate system of the bearing surface to achieve defect localization. The specific location of the defect on the bearing surface can be intuitively and clearly determined, facilitating timely targeted repair or adjustment. This greatly ensures the performance and service life of copper-based self-lubricating bearings, providing an efficient and accurate detection method for related production and application fields, reducing equipment failure risks, and improving production efficiency and product quality.

[0069] Please see Figure 2 , Figure 2 This is a schematic diagram of the structural framework of one embodiment of the copper-based self-lubricating bearing surface lubricant distribution detection device of this application. Figure 2 As shown, the copper-based self-lubricating bearing surface lubricant distribution detection device 20 includes: an acquisition module 21, a segmentation module 22, a first stitching module 23, a second stitching module 24, and a positioning module 25. The acquisition module 21 acquires an image of the bearing target surface under preset environmental conditions, including identical lighting conditions and a uniform scanning speed. The target surface is the bearing surface where the lubricant distribution needs to be inspected. The segmentation module 22 performs edge detection on the target surface image, segmenting the lubricant distribution area and defect candidate areas. The defect candidate areas include several preset defect types. The first stitching module 23, based on the defect type information, stitches together the lubricant distribution sub-regions corresponding to defects displayed across regions to generate a target detection area. The second stitching module 24 detects the texture row contour points of the target detection area, calculates the tilt angle difference between adjacent texture rows, and triggers a stitching operation when the tilt angle difference meets preset stitching conditions. The positioning module 25 segments the lubricant area, generates distribution feature data, and maps it to the three-dimensional coordinate system of the bearing surface to achieve defect positioning.

[0070] Among them, the same lighting conditions include a light source system with a cross layout, which is configured to uniformly illuminate the bearing target surface.

[0071] In some embodiments, the first stitching module stitches together the lubricant distribution sub-regions corresponding to the defects displayed across regions based on the defect type information to generate a target detection region, including: in response to the fact that the spacing between defects is less than or equal to a preset spacing threshold and the contour overlap area is greater than or equal to a preset overlap ratio, determining that it is a nested distribution and performing a stitching operation.

[0072] In some embodiments, the second stitching module detects the contour points of the texture rows in the target detection area, calculates the tilt angle difference between adjacent texture rows, and triggers a stitching operation in response to the tilt angle difference meeting a preset stitching condition, including: calculating the ratio of the intersection area to the minimum area as the overlap of the texture row region; and triggering a stitching operation in response to the overlap exceeding a preset overlap threshold.

[0073] In some embodiments, the second stitching module calculates the tilt angle difference between adjacent texture rows, including: extracting the coordinates of the edge contour points on both sides of the texture row; and calculating the tilt angle difference between adjacent texture rows based on the coordinates of the edge contour points on both sides.

[0074] In some embodiments, the second stitching module calculates the overlap of the texture row regions by the ratio of the intersection area to the minimum region area, including: extracting the contour points of the texture row regions using an edge detection algorithm; constructing corresponding polygonal geometry for each pair of adjacent texture row regions based on the extracted contour points, obtaining the intersection area by calculating the intersection between the two polygons; and performing a ratio operation between the intersection area and the smaller value of the area in each pair of adjacent texture regions to obtain the overlap.

[0075] In some embodiments, the second stitching module triggers a stitching operation in response to the tilt angle difference meeting a preset stitching condition, including: calculating the tilt angle value of each texture row region, calculating its average value as a first angle value; setting a preset threshold for the angle difference of a single texture row region, comparing the tilt angle of each texture row region with the first angle value, and generating an angle difference; and triggering the calculation of overlap degree in response to the angle difference of a certain texture row region exceeding the preset threshold, and stitching the texture row regions based on the overlap degree.

[0076] In the above scheme, by acquiring images of the bearing target surface under preset environmental conditions, the stability and consistency of image acquisition are ensured. By performing edge detection on the target surface image and segmenting the lubricant distribution area and defect candidate area, potential defect areas can be quickly and accurately identified, improving detection efficiency. Based on defect type information, relevant sub-regions are stitched together to generate the target detection area. Combining the detection of the texture line contour points of the target detection area, calculating the tilt angle difference, and performing corresponding stitching operations, the detection omission problem caused by defects displaying across regions is effectively avoided, improving the comprehensiveness and accuracy of detection. Finally, the lubricant area is segmented and the distribution feature data is generated and mapped to the three-dimensional coordinate system of the bearing surface to achieve defect localization. The specific location of the defect on the bearing surface can be intuitively and clearly determined, facilitating timely targeted repair or adjustment. This greatly ensures the performance and service life of copper-based self-lubricating bearings, providing an efficient and accurate detection method for related production and application fields, reducing equipment failure risks, and improving production efficiency and product quality.

[0077] Figure 3 This is a schematic diagram of the structural framework of one embodiment of the copper-based self-lubricating bearing surface lubricant distribution detection device of this application. The copper-based self-lubricating bearing surface lubricant distribution detection device 30 can vary considerably due to different configurations or performance characteristics. It may include one or more processors 31 and a memory 32. The processor 31 may be configured to communicate with the memory 32, executing a series of instructions stored in the memory on the copper-based self-lubricating bearing surface lubricant distribution detection device to implement the steps of the aforementioned copper-based self-lubricating bearing surface lubricant distribution detection method. Those skilled in the art will understand that… Figure 3The illustrated structure of the copper-based self-lubricating bearing surface lubricant distribution detection device does not constitute a limitation on the copper-based self-lubricating bearing surface lubricant distribution detection device provided by the present invention. It may include more or fewer components than illustrated, or combine certain components, or have different component arrangements.

[0078] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when executed on a computer, cause the computer to perform the steps of the method for detecting the distribution of lubricant on the surface of a copper-based self-lubricating bearing.

[0079] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the system, device, or unit described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0080] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0081] The above 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.

Claims

1. A method for detecting the distribution of lubricant on the surface of a copper-based self-lubricating bearing, characterized in that, The method includes: acquiring a bearing target surface image under preset environmental conditions, including identical lighting conditions and a uniform scanning speed, wherein the target surface is the surface of the bearing where the lubricant distribution needs to be inspected; performing edge detection on the target surface image to segment a lubricant distribution area and a defect candidate area, wherein the defect candidate area includes several preset defect types; based on the defect type information, stitching together the lubricant distribution sub-regions corresponding to defects displayed across regions to generate a target detection area; detecting the texture row contour points of the target detection area, calculating the tilt angle difference between adjacent texture rows, and triggering a stitching operation in response to the tilt angle difference satisfying a preset stitching condition; segmenting the lubricant area, generating distribution feature data, and mapping it to the three-dimensional coordinate system of the bearing surface to achieve defect localization; The process of detecting the texture line contour points of the target detection region, calculating the tilt angle difference between adjacent texture lines, and triggering a stitching operation in response to the tilt angle difference satisfying a preset stitching condition includes: calculating the ratio of the intersection area to the minimum region area as the overlap degree of the texture line region; and triggering a stitching operation in response to the overlap degree exceeding a preset overlap threshold. The calculation of the ratio of the intersection area to the minimum region area is the overlap degree of the texture row region, which includes: extracting the contour points of the texture row region using an edge detection algorithm; for each pair of adjacent texture row regions, constructing corresponding polygonal geometry based on the extracted contour points, obtaining the intersection area by calculating the intersection between the two polygons; and performing a ratio operation between the intersection area and the smaller value of the area in each pair of adjacent texture regions to obtain the overlap degree. The step of triggering a splicing operation in response to the tilt angle difference meeting a preset splicing condition includes: calculating the tilt angle value of each texture row region, calculating its average value as a first angle value; setting a preset threshold for the angle difference of a single texture row region, comparing the tilt angle of each texture row region with the first angle value to generate an angle difference; and triggering the calculation of the overlap degree when the angle difference of a certain texture row region exceeds the preset threshold, and splicing the texture row regions based on the overlap degree.

2. The method for detecting the distribution of lubricant on the surface of a copper-based self-lubricating bearing according to claim 1, characterized in that, The same illumination conditions include a light source system with a cross-layout configuration, the light source system being configured to uniformly illuminate the bearing target surface.

3. The method for detecting the distribution of lubricant on the surface of a copper-based self-lubricating bearing according to claim 1, characterized in that, The step of stitching together the lubricant distribution sub-regions corresponding to the defects displayed across regions based on the defect type information to generate a target detection region includes: in response to the fact that the spacing between the defects is less than or equal to a preset spacing threshold and the contour overlap area is greater than or equal to a preset overlap ratio, determining that it is a nested distribution and performing a stitching operation.

4. The method for detecting the distribution of lubricant on the surface of a copper-based self-lubricating bearing according to claim 1, characterized in that, The calculation of the tilt angle difference between adjacent texture rows includes: extracting the coordinates of the edge contour points on both sides of the texture row; and calculating the tilt angle difference between adjacent texture rows based on the coordinates of the edge contour points on both sides.

5. A device for detecting the distribution of lubricant on the surface of a copper-based self-lubricating bearing, characterized in that, The copper-based self-lubricating bearing surface lubricant distribution detection device includes: an acquisition module for acquiring an image of the bearing target surface under preset environmental conditions, including identical lighting conditions and a uniform scanning speed, wherein the target surface is the surface of the bearing where the lubricant distribution needs to be inspected; a segmentation module for performing edge detection on the target surface image to segment the lubricant distribution area and a defect candidate area, wherein the defect candidate area includes several preset defect types; a first stitching module for stitching the lubricant distribution sub-regions corresponding to defects displayed across regions based on the defect type information to generate a target detection area; a second stitching module for detecting the texture row contour points of the target detection area, calculating the tilt angle difference between adjacent texture rows, and triggering a stitching operation in response to the tilt angle difference satisfying a preset stitching condition; and a positioning module for segmenting the lubricant area, generating distribution feature data, and mapping it to the three-dimensional coordinate system of the bearing surface to achieve defect positioning. The process of detecting the texture line contour points of the target detection region, calculating the tilt angle difference between adjacent texture lines, and triggering a stitching operation in response to the tilt angle difference satisfying a preset stitching condition includes: calculating the ratio of the intersection area to the minimum region area as the overlap degree of the texture line region; and triggering a stitching operation in response to the overlap degree exceeding a preset overlap threshold. The calculation of the ratio of the intersection area to the minimum region area is the overlap degree of the texture row region, which includes: extracting the contour points of the texture row region using an edge detection algorithm; for each pair of adjacent texture row regions, constructing corresponding polygonal geometry based on the extracted contour points, obtaining the intersection area by calculating the intersection between the two polygons; and performing a ratio operation between the intersection area and the smaller value of the area in each pair of adjacent texture regions to obtain the overlap degree. The step of triggering a splicing operation in response to the tilt angle difference meeting a preset splicing condition includes: calculating the tilt angle value of each texture row region, calculating its average value as a first angle value; setting a preset threshold for the angle difference of a single texture row region, comparing the tilt angle of each texture row region with the first angle value to generate an angle difference; and triggering the calculation of the overlap degree when the angle difference of a certain texture row region exceeds the preset threshold, and splicing the texture row regions based on the overlap degree.

6. A device for detecting the distribution of lubricant on the surface of a copper-based self-lubricating bearing, characterized in that, The copper-based self-lubricating bearing surface lubricant distribution detection device includes: a memory and at least one processor, the memory storing instructions; the at least one processor invokes the instructions in the memory to cause the copper-based self-lubricating bearing surface lubricant distribution detection device to perform the steps of the copper-based self-lubricating bearing surface lubricant distribution detection method as described in any one of claims 1-4.

7. A computer-readable storage medium storing instructions thereon, characterized in that, When the instruction is processed and executed, it implements the steps of the method for detecting the distribution of lubricant on the surface of a copper-based self-lubricating bearing as described in any one of claims 1-4.

Images