Intelligent processing lubricating oil station hydraulic element data monitoring and early warning method and system

Abstract

This invention discloses an intelligent data monitoring and early warning method and system for hydraulic components in a lubrication station, relating to the field of image processing technology. The method includes: constructing an aligned image set based on a first image and a second image of the inner wall of the oil passage; constructing a baseline brightness image based on the aligned image set; generating the average deviation of all working condition blocks based on the baseline brightness image, and determining the scintillation of each image micro-block corresponding to the first image to obtain a microbubble scintillation dominant block; performing mask construction and morphological expansion processing on the microbubble scintillation dominant block to obtain an extended bubble contamination mask image; and defining a de-bubble contamination image based on the extended bubble contamination mask image. This invention improves the reliability and stability of cavitation erosion monitoring results on the inner wall of the oil passage.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and in particular to a method and system for intelligently monitoring and warning data of hydraulic components in a lubrication oil station. Background Technology

[0002] As a key basic unit in large industrial equipment and power systems, lubrication stations are widely used in metallurgy, power, chemical, and heavy machinery industries. Their core function is to pressurize, distribute, and stably supply lubricating oil through hydraulic components to ensure the safe operation of equipment under high load and high speed conditions. In the hydraulic system of a lubrication station, a complex network of oil channels is formed inside the integrated valve block. Lubricating oil repeatedly passes through the inner walls of these channels under high pressure and high speed. Over long-term operation, cavitation can easily occur in localized areas, leading to pitting, micro-pitting, and progressively expanding erosion damage on the material surface. Because the inner walls of the oil channels are in a closed space and their structure is concealed, endoscopic imaging is typically used to obtain images of their surface condition on-site, and these images are periodically collected under different operating conditions to monitor and assess the health status of the hydraulic components.

[0003] In existing technologies, when judging the cavitation erosion state of the inner wall of oil passages based on endoscopic images, the problem of brightness fluctuations caused by microbubbles in the oil changing with operating conditions is commonly encountered. Such brightness changes appear in the image as alternating enhancement or weakening of local areas under different operating conditions, which can easily mask the true metal surface condition. This makes it difficult for methods based on a single operating condition or direct brightness analysis to distinguish between image fluctuations caused by changes in fluid state and structural changes caused by material surface degradation. As a result, there is a risk of misjudgment or omission in the monitoring results. There is a lack of an effective mechanism for systematically identifying and eliminating brightness changes related to operating conditions. It is difficult to stably and accurately reflect the true cavitation erosion development of the inner wall of the oil passages under complex operating conditions, which limits the reliability and practicality of monitoring and early warning of hydraulic components in lubrication stations. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies that limit the reliability and practicality of monitoring and early warning of hydraulic components in lubrication stations, and to propose an intelligent data processing method and system for monitoring and early warning of hydraulic components in lubrication stations.

[0005] To address the problems existing in the prior art, the present invention adopts the following technical solution: A method for intelligently monitoring and early warning of hydraulic component data in a lubrication oil station includes: Construct an aligned image set based on the first and second images of the inner wall of the oil passage; Construct a baseline brightness image based on an aligned image set; The average deviation of the block under all operating conditions is generated based on the baseline brightness image, and the flicker is determined for each microblock corresponding to the first image to obtain the microbubble flicker dominant block; A mask is constructed and morphological dilation is performed on the microbubble scintillation dominant block to obtain an extended bubble contamination mask image; a de-bubble contamination image is defined based on the extended bubble contamination mask image; The cavitation erosion state of the inner wall of the oil passage is monitored and warned by debubbling contamination images.

[0006] Preferably, the specific steps for constructing the aligned image set are as follows: Positioning the integrated valve block of the hydraulic system in the lubrication station; With the position of the inner wall of the oil passage of the integrated valve block remaining unchanged, the hydraulic system is sequentially set to different hydraulic conditions; wherein, the different hydraulic conditions include at least: a first hydraulic condition and a second hydraulic condition; Under the first hydraulic condition, acquire the first image of the inner wall of the oil passage; Under the second hydraulic condition, a second image of the inner wall of the oil passage is acquired; Use the first image as the reference image; Determine the spatial transformation relationship between the second image and the reference image; Based on the spatial transformation relationship, the second image is subjected to coordinate transformation to obtain the second aligned image; Construct an alignment image set based on the reference image and the second alignment image.

[0007] Preferably, constructing a baseline brightness image based on an aligned image set includes: Get the grayscale value at pixel coordinates for each aligned image in the aligned image set; The median of the grayscale values ​​is calculated to obtain the baseline grayscale value; A baseline brightness image is constructed based on all baseline grayscale values.

[0008] Preferably, generating the average deviation of the entire operating condition block based on the baseline brightness image includes: Subtract the baseline gray value of the baseline brightness image from the gray value of the aligned image to obtain the brightness deviation value; The observation area of ​​the first image is divided into multiple image micro-blocks; Under each hydraulic condition, the brightness deviation values ​​at all pixel coordinates in the image micro-block are summed to obtain the total brightness deviation. Divide the sum of brightness deviations by the total number of pixels in the image micro-blocks to obtain the average block deviation. The average block deviation of the image micro-block under all hydraulic conditions is calculated to obtain the average block deviation under all conditions.

[0009] Preferably, the specific steps for obtaining the microbubble scintillation dominant block are as follows: Based on the numerical comparison between the average block deviation and the average block deviation under all working conditions, the image micro-blocks are marked with status to obtain the marking results. Based on the labeling results of the image micro-blocks under all hydraulic conditions, a sequence of operating conditions for each image micro-block is generated; By analyzing the flickering of each image micro-block using the operating condition sequence, the dominant microbubble flickering block is obtained.

[0010] Preferably, the specific steps for obtaining the extended bubble contamination mask image are as follows: Construct a mask image based on the first image; Set the mask value of all pixel coordinates in the mask image to 0; In the mask image, the mask value corresponding to all pixel coordinates belonging to the microbubble scintillation dominant block is set from 0 to 1 to obtain the first bubble contamination mask image; Morphological dilation processing is performed on the first bubble contamination mask image to obtain an expanded bubble contamination mask image.

[0011] Preferably, the specific steps for defining a de-bubble contamination image are as follows: Construct an initial grayscale image based on the first image; When the mask value of the extended bubble contamination mask image at the pixel coordinates is 0, the pixel gray value of the baseline brightness image at the pixel coordinates is used as the pixel gray value of the initial grayscale image at the pixel coordinates. When the mask value of the expanded bubble contamination mask image at the pixel coordinates is 1, the pixel gray value of the initial grayscale image at the pixel coordinates is set to an invalid value. Define the initial image as a bubble-removed contamination image.

[0012] Preferably, the specific steps for monitoring and early warning of cavitation erosion are as follows: Determine the grayscale gradient magnitude of image micro-patch in the de-bubbling contamination image; Sum the magnitudes of all grayscale gradients to obtain the total grayscale gradient. Divide the sum of grayscale gradients by the total number of pixels in the image micro-block to obtain the block roughness value; The block roughness values ​​of all image micro-blocks are averaged to obtain the global roughness mean. Based on the numerical comparison results of the global roughness mean and the block roughness value, roughness level is marked for each image micro-block to obtain the erosion risk marking result; A cavitation erosion risk distribution map is generated based on the erosion risk labeling results; The cavitation erosion status of the inner wall of the oil passage is monitored and warned by using a cavitation erosion risk distribution map.

[0013] To address the aforementioned problems, the present invention also provides an intelligent data monitoring and early warning system for hydraulic components in a lubrication oil station, the system comprising: The alignment module is used to construct an aligned image set based on the first and second images of the inner wall of the oil passage; The baseline module is used to construct a baseline brightness image based on an aligned set of images; The flicker module is used to generate the average deviation of the block under all operating conditions based on the baseline brightness image, and to perform flicker determination on each image microblock corresponding to the first image to obtain the microbubble flicker dominant block; The mask module is used to construct a mask and perform morphological dilation processing on the microbubble scintillation dominant block to obtain an extended bubble contamination mask image; a de-bubble contamination image is defined based on the extended bubble contamination mask image; The early warning module is used to monitor and provide early warning of the cavitation erosion state of the inner wall of the oil passage through debubbling contamination images.

[0014] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention acquires first and second images of the same oil passage inner wall location under different hydraulic operating conditions, and constructs an aligned image set based on spatial transformation relationships. This ensures that the images under different operating conditions correspond to the same actual position of the oil passage inner wall at the same pixel coordinates. As a result, all subsequent grayscale comparisons, difference calculations, and block statistics based on pixel coordinates are established on the same spatial reference system, avoiding regional mismatches caused by slight displacement of the endoscope, viewing angle deviation, or scale differences. This makes the operating condition difference analysis comparable and consistent from the source, thus providing a reliable data foundation for the stable assessment of the oil passage inner wall condition.

[0015] 2. This invention constructs a baseline brightness image based on an aligned image set, and further generates the average deviation value of the blocks under all operating conditions, completes the state marking of image micro-blocks and the generation of operating condition state sequences, and then performs a flicker determination on each image micro-block to obtain the microbubble flicker dominant block. This enables the identification of the alternating bright and dark areas caused by microbubbles in the oil by the brightness deviation pattern across operating conditions, under the condition of acquiring only one image for each operating condition, and distinguishes such areas from the changes in the actual metal surface structure. This avoids the enhanced specular reflection or refractive changes caused by microbubbles being misjudged as increased surface roughness, and also avoids the omission of the actual early erosion area due to the coverage of bright and dark fluctuations, thus improving the reliability and stability of the monitoring results of cavitation erosion on the inner wall of the oil passage. Attached Figure Description

[0016] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings: Figure 1 This is a flowchart illustrating an intelligent data monitoring and early warning method for hydraulic components in a lubrication station, provided as an embodiment of the present invention. Detailed Implementation

[0017] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0018] This embodiment provides a method for intelligently monitoring and issuing early warnings for hydraulic component data in a lubrication oil station. (See also...) Figure 1 Specifically, including: S1. Construct an aligned image set based on the first and second images of the inner wall of the oil passage; In an embodiment of the present invention, the specific steps for constructing the aligned image set are as follows: Positioning the integrated valve block of the hydraulic system in the lubrication station; The hydraulic system in a lubrication station refers to the system structure used to pressurize, distribute, and control lubricating oil. The integrated valve block is a hydraulic control component installed in the hydraulic system, which has multiple oil passages inside to realize the switching of lubricating oil flow direction and pressure regulation.

[0019] Start the hydraulic system of the lubrication station and shut it down to ensure operational safety. Then, trace the main pipeline of the lubrication station's hydraulic system, starting from the outlet of the hydraulic pump, and check the core control nodes where the pipelines intersect. Identify the block-shaped structure with multiple hydraulic valves installed and multiple pipeline interfaces distributed on its surface. This structure is the candidate for the integrated valve block. Next, observe the external morphological characteristics of the candidate structure to confirm that it has multiple oil passages connecting the various hydraulic valves and can realize the function of switching the flow direction and regulating the pressure of lubricating oil through the internal oil passages. At the same time, verify that the structure is directly connected to the core components of the hydraulic system, such as the actuators and filters, through pipelines. Further, by consulting the assembly drawings and technical documents of the lubrication station's hydraulic system, compare the model, size, installation position, and other parameters of the candidate structure with the integrated valve block information recorded in the documents. If all parameters match, the integrated valve block is located. If there are parameter mismatches, return to the pipeline tracing step to check again until the integrated valve block that meets all the characteristics and parameter requirements is accurately located.

[0020] With the position of the inner wall of the oil passage of the integrated valve block remaining unchanged, the hydraulic system is sequentially set to different hydraulic conditions; wherein, the different hydraulic conditions include at least: a first hydraulic condition and a second hydraulic condition; First, confirm the observation position on the inner wall of the oil passage inside the integrated valve block. By fixing the geometric installation position and observation direction of the endoscope probe and adjusting the light source to maintain stable light intensity, ensure that the observation target area on the inner wall of the oil passage is always at the same monitoring angle and under consistent environmental conditions. Then, start the control unit of the hydraulic system. By adjusting the core control components such as the system pressure regulation component, flow control component, and oil flow path switching mechanism, gradually change the operating parameters of the hydraulic system. First, adjust the hydraulic system to the first hydraulic condition and continuously monitor the system pressure value, lubricating oil flow rate, and oil flow state. After all parameters tend to stabilize and the lubricating oil flow in the oil passage no longer fluctuates significantly, keep all settings related to the observation position unchanged and continue to adjust the pressure, flow rate, or flow path related parameters through the control unit to switch the hydraulic system to the second hydraulic condition, which is significantly different from the first hydraulic condition. Monitor the system operating parameters again until they stabilize, ensuring that the first and second hydraulic conditions correspond to different pressure levels, flow states, or oil flow modes, thereby causing the lubricating oil at the same physical location in the oil passage to exhibit different flow characteristics and pressure conditions.

[0021] The inner wall of the oil passage refers to the solid surface inside the integrated valve block where the oil passage directly contacts the lubricating oil; the hydraulic condition refers to the working state of the hydraulic system during operation, which is determined by pressure, flow rate and flow state. Different hydraulic conditions correspond to differences in the flow characteristics and pressure conditions of the lubricating oil in the oil passage.

[0022] The first hydraulic condition and the second hydraulic condition refer to two different working states set sequentially in the hydraulic system of the lubrication station, under the condition that the observation position of the inner wall of the oil passage of the integrated valve block remains unchanged. The first hydraulic condition and the second hydraulic condition correspond to different pressure levels, flow states or oil flow patterns, so that the lubricating oil in the oil passage exhibits different flow and force states at the same physical location, thus providing a basis for subsequent analysis of the differences in image features of the inner wall of the same oil passage under different hydraulic working states.

[0023] Under the first hydraulic condition, acquire the first image of the inner wall of the oil passage; After the hydraulic system is running stably in the first hydraulic condition, the system pressure, lubricating oil flow rate, and oil flow status are continuously monitored. When all parameters remain stable and without significant fluctuations, the probe of the endoscope imaging system is slowly inserted into the target oil passage of the integrated valve block. By adjusting the geometric installation position and observation angle of the probe, the observation area of ​​the inner wall of the oil passage is fully presented in the center of the imaging field of view. At the same time, the illumination intensity and angle of the light source are adjusted to ensure that the light uniformly covers the entire observation area and avoid strong reflections or local shadows that affect the image clarity. Then, the image acquisition function of the endoscope is activated to take pictures of the surface condition of the inner wall of the oil passage under the current first hydraulic condition, generating image data that can clearly reflect the texture of the solid surface of the inner wall of the oil passage, potential micro-pits, and other surface features. This image data is stored in the system's designated storage unit and defined as the first image.

[0024] Under the second hydraulic condition, a second image of the inner wall of the oil passage is acquired; After the hydraulic system completes the acquisition of the first image under the first hydraulic condition, the geometric installation position, observation angle, and illumination intensity and angle of the endoscope probe remain unchanged. The relevant operating parameters are adjusted through the hydraulic system control unit to switch the system to the second hydraulic condition. The system pressure, lubricating oil flow rate, and oil flow status are continuously monitored. When all parameters are stable and without significant fluctuations, it is confirmed that the observation area of ​​the inner wall of the oil passage is still completely presented in the center of the endoscope's imaging field of view, and the light evenly covers the area without producing new strong reflections or local shadows. The image acquisition function of the endoscope is activated to capture the surface condition of the inner wall of the oil passage under the current second hydraulic condition, generating image data that can clearly reflect the texture of the solid surface of the inner wall of the oil passage, potential micro-pits, and other surface features under this condition. The image data is stored in the system's designated storage unit and defined as the second image.

[0025] The first image refers to image data reflecting the surface condition of the inner wall of the oil passage, acquired under the first hydraulic condition. The second image refers to image data reflecting the surface condition of the inner wall of the same oil passage, acquired under the second hydraulic condition.

[0026] Use the first image as the reference image; Determine the spatial transformation relationship between the second image and the reference image; Feature extraction is performed on the reference image and the second image respectively. An edge detection algorithm is used to identify structural feature points with stable recognizability, such as the inherent contour edges of the inner wall of the oil passage and naturally formed small protrusions or depressions on the surface. The extracted feature points are filtered to remove invalid feature points that are blurry, isolated, or have excessive gray value fluctuations, and retain valid feature points with clear contours that can be stably presented in both images. Then, a feature point matching algorithm is used to calculate the gray-level similarity and geometric distance between valid feature points to establish an initial correspondence between valid feature points in the reference image and the second image. Then, a random sampling consensus algorithm is used to iteratively filter and remove abnormal feature points with mismatches, retaining the correct feature point correspondence that meets the geometric constraints. Based on the correct feature point correspondence, an overdetermined system of equations is constructed. The least squares method is used to solve the system of equations to obtain transformation parameters such as translation, rotation angle, scaling ratio, and shearing coefficient. These parameters together constitute a spatial transformation relationship that can accurately describe the geometric mapping relationship between the pixel positions in the second image and the corresponding pixel positions in the reference image, ensuring that the spatial transformation relationship can accurately compensate for the geometric deviations caused by small displacements, angular shifts, or scale differences that may occur in the two images during the shooting process.

[0027] The reference image refers to the image data selected from the first image and the second image as a spatial coordinate reference; the spatial transformation relationship refers to the mapping relationship used to describe the geometric correspondence between the pixel positions in the second image and the corresponding pixel positions in the reference image.

[0028] Based on the spatial transformation relationship, the second image is subjected to coordinate transformation to obtain the second aligned image; Obtain the number of pixel rows and columns of the reference image. Set the pixel size of the second aligned image to be completely consistent with the reference image to ensure that there is no scale difference in the coordinate system. Then, calculate the inverse transformation matrix corresponding to the spatial transformation relationship. Establish a mapping relationship between the pixel coordinates of the second aligned image and the pixel coordinates of the original second image through this inverse transformation matrix. Then, traverse each target pixel coordinate of the second aligned image row by row and column by column. For each target pixel coordinate, substitute it into the inverse transformation matrix to calculate its corresponding floating-point coordinate in the original second image. Perform boundary verification on the floating-point coordinate. If the row value or column value of the floating-point coordinate exceeds the pixel size of the original second image, the corresponding floating-point coordinate is not verified. If the target pixel's grayscale value is within the specified range, the grayscale value of the current target pixel is set to invalid. If it is within the valid range, the four nearest integer pixel coordinates around the floating-point coordinates are selected. Weights are assigned based on the distance between the four integer pixel coordinates and the floating-point coordinates. The grayscale value of the target pixel is calculated by weighted summation. The weight assignment follows the principle that the closer the pixel is to the floating-point coordinate, the greater the weight. After calculating the grayscale values ​​of all target pixels in sequence, a second aligned image is formed that is in the same pixel coordinate system as the reference image. Each pixel in this image points precisely to the same physical location on the inner wall of the oil passage as the corresponding pixel in the reference image, achieving a one-to-one correspondence in spatial location.

[0029] Coordinate transformation refers to adjusting the position of each pixel in the second image according to the spatial transformation relationship, so that it is in the same pixel coordinate system as the reference image; the second aligned image refers to the second image that corresponds one-to-one with the reference image in spatial position after coordinate transformation.

[0030] Construct an alignment image set based on the reference image and the second alignment image.

[0031] The pixel dimension information of the reference image and the second aligned image is read separately to confirm that the number of pixel rows and columns of the two images are completely consistent, ensuring that they are in the same pixel coordinate system and have no scale difference. Then, a unified image storage index rule is established to assign unique working condition identification information to the reference image and the second aligned image. This identification information corresponds one-to-one with the hydraulic working condition at the time of image acquisition. At the same time, the acquisition time, endoscope probe position parameters and other related information of each image are recorded. Then, the reference image and the second aligned image are written into the same data set in a preset storage format. The mapping relationship between the pixel coordinates within the data set and the actual physical position of the inner wall of the oil passage is established to ensure that the same pixel coordinate of each image in the data set accurately corresponds to the same actual area of ​​the inner wall of the oil passage. Finally, the integrity of the data set is checked to confirm that the reference image and the second aligned image have been completely stored and the related information is accurate, forming an aligned image set composed of the reference image and the second aligned image.

[0032] An aligned image set refers to an image data set consisting of a reference image and a second aligned image, where each image in the set corresponds to the same actual position on the inner wall of the oil passage at the same pixel coordinate.

[0033] S2. Construct a baseline brightness image based on the aligned image set; In an embodiment of the present invention, constructing a baseline brightness image based on an aligned image set includes: Get the grayscale value at pixel coordinates for each aligned image in the aligned image set; An aligned image refers to a single image in the set of aligned images whose pixel coordinates are in the same coordinate system as the reference image; pixel coordinates refer to a two-dimensional position identifier used to uniquely determine the position of a pixel in the image; grayscale value refers to the numerical information reflecting the brightness intensity of the inner wall surface of the oil passage at the corresponding pixel position.

[0034] The pixel dimension information of each aligned image is read one by one to determine the number of pixel rows and columns in each image to define the effective pixel coordinate range. Then, each aligned image in the aligned image set is selected in a preset order. For the currently selected aligned image, all pixel coordinates are traversed one by one in the order from left to right and from top to bottom. Each pixel coordinate is uniquely identified in the image by a two-dimensional position identifier. During the traversal, the image grayscale information reading algorithm is used to extract the numerical information reflecting the brightness of the inner wall surface of the oil passage at each pixel coordinate. At the same time, the aligned image identifier and the row and column numbers of the pixel coordinates are recorded to ensure that each grayscale value is uniquely associated with the corresponding aligned image and pixel position. During the traversal, it is simultaneously checked whether each pixel coordinate is within the effective pixel range of the current aligned image. Invalid coordinates that are out of range are eliminated to ensure that all extracted grayscale values ​​are valid data. The traversal of all aligned images and the extraction of grayscale values ​​at each pixel coordinate are completed in sequence, and finally, the complete grayscale value information of each aligned image in the aligned image set at the corresponding pixel coordinate is obtained.

[0035] The median of the grayscale values ​​is calculated to obtain the baseline grayscale value; For each pixel coordinate, first collect all gray values ​​from all aligned images at that coordinate. Sort the collected gray values ​​in numerical order. If the number of collected gray values ​​is odd, select the gray value in the middle position after sorting as the baseline gray value at that pixel coordinate. If the number of collected gray values ​​is even, select the average of the gray values ​​in the middle two positions after sorting as the baseline gray value at that pixel coordinate. This value can characterize the relatively stable brightness level at the corresponding position on the inner wall of the oil passage under different hydraulic conditions. The median calculation is completed for all pixel coordinates in sequence to obtain the baseline gray value corresponding to each pixel coordinate.

[0036] Median calculation refers to the process of sorting the gray values ​​from multiple aligned images at the same pixel coordinate and selecting the value at the center position; the baseline gray value refers to the gray value obtained by the median calculation, which is used to characterize the relatively stable brightness level of the inner wall of the oil passage under different hydraulic conditions.

[0037] A baseline brightness image is constructed based on all baseline grayscale values.

[0038] A baseline brightness image is an image composed of the baseline gray values ​​at all pixel coordinates, used to reflect the brightness distribution of the stable surface structure of the inner wall of the oil passage under various hydraulic conditions.

[0039] Determine the pixel size of the baseline brightness image, ensuring that its number of pixel rows and columns are exactly the same as those of the images in the aligned image set, and that both are in the same pixel coordinate system. Next, create a blank image frame that matches this size, and define the two-dimensional coordinate identifier corresponding to each pixel position in the frame. Then, fill the corresponding position of the blank image frame with the baseline gray value corresponding to each pixel coordinate one by one, in order from left to right and from top to bottom, ensuring that each baseline gray value corresponds precisely to the pixel coordinate without any positional offset. After filling, perform gray value continuity verification on the image, and finally form a baseline brightness image composed of the baseline gray values ​​of all pixel coordinates.

[0040] S3. Generate the average deviation of the block under all operating conditions based on the baseline brightness image, and perform flicker determination on each micro-block corresponding to the first image to obtain the microbubble flicker dominant block; In an embodiment of the present invention, generating the average deviation of the entire operating condition block based on the baseline brightness image includes: Subtract the baseline gray value of the baseline brightness image from the gray value of the aligned image to obtain the brightness deviation value; The brightness deviation value refers to the difference between the gray value of the aligned image at a certain pixel coordinate and the baseline gray value of the baseline brightness image at the same pixel coordinate. It is used to reflect the brightness change of that position relative to the steady state under specific hydraulic conditions.

[0041] The observation area of ​​the first image is divided into multiple image micro-blocks; The observation area refers to the image region in the first image used to analyze the state of the inner wall of the oil passage; the image micro-block refers to the local sub-region obtained by dividing the image within the observation area of ​​the inner wall of the oil passage. Each image micro-block consists of a set of continuous pixel coordinates and corresponds to a small actual surface area on the inner wall of the oil passage, used to statistically analyze the brightness variation characteristics of the local area under different hydraulic conditions.

[0042] First, the first image is scanned as a whole to identify the effective image range that clearly reflects the surface structure of the inner wall of the oil passage without blurring, invalid grayscale areas, or redundant background information. This range is defined as the observation area. Then, the pixel coordinate detection method is used to determine the start and end positions of the rows, columns, and pixels in the observation area, thus defining the complete pixel coordinate boundary of the observation area. Next, based on the total number of rows and columns of pixels in the observation area, the segmentation interval is set along the horizontal and vertical directions according to the principle of uniform segmentation. The segmentation interval must ensure that each local sub-region formed after the division is of the same size and does not overlap. The horizontal division is performed from the start position of the row to the end position of the row according to the segmentation interval, and the vertical division is performed from the start position of the column to the end position of the column according to the segmentation interval. Finally, multiple regular rectangular local sub-regions are formed. Each local sub-region consists of a set of continuous and unique pixel coordinates, and each local sub-region corresponds to a continuous actual surface area on the inner wall of the oil passage. These rectangular local sub-regions are the image micro-blocks.

[0043] Under each hydraulic condition, the brightness deviation values ​​at all pixel coordinates in the image micro-block are summed to obtain the total brightness deviation. Divide the sum of brightness deviations by the total number of pixels in the image micro-blocks to obtain the average block deviation. The sum of brightness deviations refers to the value obtained by accumulating the brightness deviations at the coordinates of all pixels belonging to the same image micro-block under a certain hydraulic condition; the total number of pixels in the image micro-block refers to the total number of pixels contained in the image micro-block; the average block deviation indicates the degree to which the overall brightness of the image micro-block deviates from the baseline state under a certain hydraulic condition.

[0044] First, identify all the divided image micro-blocks and the brightness deviation value corresponding to each pixel coordinate under each hydraulic condition. Then, select each hydraulic condition one by one in sequence. For the currently selected hydraulic condition, iterate through all image micro-blocks one by one. For each image micro-block, first determine all the continuous pixel coordinates it contains, collect the brightness deviation value corresponding to all pixel coordinates in the image micro-block, and accumulate these brightness deviation values ​​to obtain the total brightness deviation of the image micro-block under the current hydraulic condition. At the same time, count the total number of pixels contained in the image micro-block, i.e., the total number of pixels. Then, divide the calculated total brightness deviation by the total number of pixels in the image micro-block to obtain the average block deviation of the image micro-block under the current hydraulic condition. After calculating the total brightness deviation and average block deviation of all image micro-blocks under the current hydraulic condition, select the next hydraulic condition and repeat the above operation until the total brightness deviation and average block deviation of each image micro-block under all hydraulic conditions have been calculated.

[0045] The average block deviation of the image micro-block under all hydraulic conditions is calculated to obtain the average block deviation under all conditions.

[0046] The average block deviation under all operating conditions refers to the value obtained by further averaging the average block deviations calculated under different hydraulic conditions for the same image micro-block. This value is used to characterize the overall deviation of the image micro-block from the baseline brightness state under the combined effect of multiple hydraulic conditions.

[0047] Each image micro-block is selected sequentially. For the currently selected image micro-block, the average block deviation obtained under all hydraulic conditions is collected. The total number of collected average block deviation values ​​is the total number of hydraulic conditions. These average block deviation values ​​are accumulated sequentially to obtain the sum of the average block deviation values ​​of the image micro-block under all hydraulic conditions. Then, the sum is divided by the total number of hydraulic conditions to obtain the average block deviation of the image micro-block under all conditions. This value can accurately characterize the overall deviation of the image micro-block from the baseline brightness state under the combined effect of multiple hydraulic conditions.

[0048] In an embodiment of the present invention, the specific steps for obtaining the microbubble scintillation dominant block are as follows: Based on the numerical comparison between the average block deviation and the average block deviation under all working conditions, the image micro-blocks are marked with status to obtain the marking results. State labeling refers to the classification identifier assigned to an image micro-block under a certain hydraulic condition based on the numerical relationship between the average block deviation and the average block deviation under all working conditions. It is used to describe the brightness change state of the micro-block under that working condition. Labeling result refers to the set of state labels assigned to the image micro-block under various hydraulic conditions during the analysis of the image micro-block, based on the numerical relationship between the average block deviation under different hydraulic conditions and the average block deviation under all working conditions. It is used to reflect the brightness change characteristics of the image micro-block under different working conditions.

[0049] For each segmented image micro-block, firstly, the average block deviation calculated for that micro-block under all hydraulic conditions is collected. The total number of these average block deviations is counted, and the sum of all average block deviations is calculated and divided by the total number to obtain the average block deviation for that micro-block under all conditions. Next, the maximum and minimum values ​​among all average block deviations for that micro-block are calculated to determine the overall range of variation of the average block deviation. Simultaneously, the amplitude of this overall range, i.e., the difference between the maximum and minimum values, is calculated. Then, each hydraulic condition is selected sequentially, and the average block deviation of that micro-block under the current hydraulic condition is extracted. The difference between this average block deviation and the average block deviation for all conditions is calculated, and a judgment is made. The absolute value of the difference is determined in relation to the overall range of variation. If the ratio is small and the difference is in the middle range of the overall range of variation, the image micro-block is marked as stable under the current hydraulic condition. If the average deviation of the block is greater than the average deviation of all blocks under all conditions and the ratio of the absolute value of the difference to the amplitude value is large, showing a significant increasing trend, it is marked as bright. If the average deviation of the block is less than the average deviation of all blocks under all conditions and the ratio of the absolute value of the difference to the amplitude value is large, showing a significant decreasing trend, it is marked as dark. After marking all hydraulic conditions in sequence, the state labels of the image micro-block under each hydraulic condition are integrated to form a marking result containing the corresponding state information for each condition.

[0050] Based on the labeling results of the image micro-blocks under all hydraulic conditions, a sequence of operating conditions for each image micro-block is generated; The operating condition state sequence refers to a data sequence formed by arranging the corresponding state markers of the same image micro-patch under all hydraulic operating conditions in the order of hydraulic operating conditions. It is used to reflect the brightness change pattern of the image micro-patch under different hydraulic operating conditions.

[0051] For each image micro-block, the marking results obtained under different hydraulic conditions are collected. The marking results are derived from the state judgment formed by comparing the average block deviation of the image micro-block under each hydraulic condition with the average block deviation of the entire condition. Then, according to the actual setting order of the hydraulic conditions, the state marks of the image micro-block under each hydraulic condition are arranged sequentially, so that the state marks of the same image micro-block under different hydraulic conditions form an ordered data combination. This ordered data combination is the working condition state sequence of the image micro-block. Each state mark in the working condition state sequence corresponds to a specific hydraulic condition, thus fully reflecting the overall evolution process of the brightness change state of the image micro-block under different hydraulic conditions.

[0052] By analyzing the flickering of each image micro-block using the operating condition sequence, the dominant microbubble flickering block is obtained.

[0053] The microbubble scintillation dominant block refers to the image micro-block determined after further judgment based on the working condition sequence formed by the marking results. The brightness state of the image micro-block under different hydraulic working conditions shows obvious variation characteristics, indicating that the image features of the oil passage inner wall area corresponding to the image micro-block are mainly dominated by the brightness scintillation phenomenon caused by microbubbles in the oil.

[0054] For each image micro-block with a generated operating condition sequence, firstly, all state markers in the sequence are thoroughly reviewed to clarify the hydraulic operating condition corresponding to each state marker and the brightness change characteristics under that condition. Then, the state markers in the sequence are checked one by one to determine whether there are both bright and dark state markers. If both exist, it indicates that the image micro-block exhibits obvious brightness reversal under different hydraulic operating conditions, which conforms to the core characteristics of microbubble scintillation changing with operating conditions. It is directly identified as a microbubble scintillation dominant block. If all state markers in the sequence are bright, the baseline brightness image is retrieved, and the baseline grayscale values ​​of the image micro-block are compared with those of the surrounding adjacent image micro-blocks to observe whether there is a significant brightness discontinuity. If there is an isolated bright appearance that has no corresponding relationship with the edge of the surrounding metal structure, it indicates that the bright feature of this area is caused by a stable gas nucleus or abnormal interface reflection. It is also identified as a microbubble scintillation dominant block. If there are only stable state markers in the sequence, or although there is a single bright or dark state, it is continuous with the brightness of the surrounding baseline. It is identified as a non-microbubble scintillation dominant block. The scintillation determination of all image micro-blocks is completed in sequence.

[0055] Microbubble scintillation refers to the image formed by changes in optical reflection and refraction conditions caused by the morphological changes, volume changes, and slight positional movements of tiny bubbles in the oil under different hydraulic conditions in a local area of ​​the inner wall of the integrated valve block. This image appears as a fine, fragmented bright and dark texture with alternating brightness increases and decreases at the same spatial location as the hydraulic conditions change. It usually presents as a granular or mottled distribution and covers the original texture of the metal surface. This type of texture is highly sensitive to hydraulic conditions. Its brightness and darkness changes are not caused by changes in the structure of the metal material itself, but are dominated by the brightness scintillation effect induced by the microbubble interface.

[0056] Because in the early pitting or micro-pit areas of the oil passage inner wall of the integrated valve block, the microbubbles in the oil undergo changes in interface morphology and position under different hydraulic conditions, resulting in significant changes in local reflection and refraction conditions. This causes the same local area to exhibit alternating brightness increases and decreases under different conditions. This brightness reversal caused by changes in operating conditions is not caused by changes in the metal material structure itself, but is dominated by optical effects caused by microbubbles. If the image is not segmented and the brightness change characteristics related to operating conditions are not analyzed block by block, and bright spots or roughness are directly used as the basis for damage, it is easy to misjudge the high brightness caused by microbubbles as surface damage, or to miss the real early erosion area because it is covered by the flickering of brightness and darkness. Therefore, by performing flicker judgment on the brightness state changes of each image micro-block under different hydraulic conditions, and identifying the image micro-blocks dominated by the microbubble flickering effect, it is possible to distinguish these areas from the real metal structure texture in subsequent processing, thus providing a reliable basis for removing bubble contamination and accurately assessing the cavitation erosion state of the oil passage inner wall.

[0057] S4. Perform mask construction and morphological dilation processing on the microbubble scintillation dominant block to obtain an extended bubble contamination mask image; define a de-bubble contamination image based on the extended bubble contamination mask image; In an embodiment of the present invention, the specific steps for obtaining the extended bubble contamination mask image are as follows: Construct a mask image based on the first image; Obtain the complete pixel dimension information of the first image, including the total number of rows and columns of pixels in the image, and determine the range of pixel coordinates of the first image. Then, based on the pixel dimension information, create a brand new blank auxiliary image, ensuring that the total number of rows and columns of pixels in the blank image is completely consistent with the first image, so that each pixel coordinate of the blank image can correspond precisely to the pixel coordinates of the first image. The blank auxiliary image that matches the size of the first image and corresponds one-to-one with the coordinates is the mask image.

[0058] Set the mask value of all pixel coordinates in the mask image to 0; Define the number of pixel rows, columns, and the complete range of pixel coordinates in the mask image. Then, traverse all pixel coordinates of the mask image one by one from left to right and from top to bottom. For each traversed pixel coordinate, set its corresponding identifier value, i.e., the mask value, to 0. This value indicates that the current pixel position has not yet been marked as a bubble contamination area.

[0059] A mask image is an auxiliary image that corresponds one-to-one with the first image in pixel coordinates. Each pixel position is used to identify whether the position is affected by the microbubble flickering effect. A mask value is a numerical value set in the mask image to indicate the state of the corresponding pixel position. A mask value of zero indicates that the pixel position is not marked as a bubble contamination area, and a mask value of one indicates that the pixel position is marked as a bubble contamination area.

[0060] In the mask image, the mask value corresponding to all pixel coordinates belonging to the microbubble scintillation dominant block is set from 0 to 1 to obtain the first bubble contamination mask image; The first bubble contamination mask image refers to the image obtained by setting the mask value corresponding to the pixel position belonging to the microbubble scintillation dominant block to one in the mask image, and is used to identify the initially determined bubble contamination area.

[0061] The range of continuous pixel coordinates contained in each microbubble scintillation dominant block is clearly defined, including the start and end positions of the pixel rows, the start and end positions of the columns, and the end positions of the columns. This ensures accurate acquisition of the specific coordinate information of all pixels within each dominant block. Subsequently, each microbubble scintillation dominant block is selected sequentially. For the currently selected dominant block, all pixel coordinates contained therein are traversed in order from left to right and from top to bottom. During the traversal, it is simultaneously checked whether each pixel coordinate is within the effective pixel range of the mask image to avoid invalid operations where the coordinates exceed the boundary. For pixel coordinates within the effective range, their corresponding mask values ​​are modified from the initial 0 to 1. After adjusting the mask values ​​of all pixel coordinates within all microbubble scintillation dominant blocks, an image is formed on the basis of the mask image, with only the pixel positions corresponding to the microbubble scintillation dominant blocks marked as 1. This image is the first bubble contamination mask image, which can accurately identify the initially determined bubble contamination area.

[0062] Because the brightness variation of the oil passage inner wall region corresponding to the microbubble scintillation dominant block under different hydraulic conditions is mainly dominated by the reflection and refraction effects caused by microbubbles in the oil, the image brightness in this type of region is unstable and fluctuates significantly with changes in operating conditions, and can no longer reliably reflect the true structural state of the metal surface. By marking the pixel coordinates corresponding to this type of region as bubble contamination areas in the mask image, it is possible to clearly distinguish between the areas affected by microbubbles and those not affected by microbubbles in subsequent image processing. This allows subsequent debubbling and cavitation erosion analysis to specifically avoid or process these pixel positions dominated by the microbubble scintillation effect.

[0063] Morphological dilation processing is performed on the first bubble contamination mask image to obtain an expanded bubble contamination mask image.

[0064] Morphological dilation refers to an image processing operation that spatially expands the marked area in the first bubble contamination mask image, so that the marked area expands spatially to cover adjacent areas that may be affected by microbubbles but have not yet been directly marked.

[0065] An extended bubble contamination mask image refers to a mask image formed by spatially expanding the marked bubble contamination areas based on the first bubble contamination mask image. The marked pixel positions in this image include not only the areas corresponding to the microbubble scintillation dominant blocks, but also the spatially adjacent areas that may be indirectly affected by the microbubble scintillation effect. This is used to more completely cover the range of microbubble-affected areas in the inner wall of the oil passage in subsequent image processing, thereby avoiding interference from the microbubble scintillation effect on the analysis of the real metal surface texture.

[0066] Based on the characteristics of the microbubble scintillation effect in the inner wall of the oil passage, a suitable structural element is selected. The size and shape of the structural element are determined according to the range of adjacent pixels where the microbubble may diffuse. If the microbubble effect is isotropically distributed, a circular structural element can be selected; if it easily diffuses along the texture direction, a rectangular or cross-shaped structural element that matches the texture direction is selected. The size of the structural element should be reasonably set in combination with the pixel span of the indirect influence of microbubbles in actual detection. Then, the selected structural element is traversed pixel by pixel on the first bubble contamination mask image in a left-to-right, top-to-bottom order. For each currently traversed pixel coordinate, the structural element covers the pixel and its surrounding adjacent pixels to form a neighborhood range. It is then determined whether there is at least one such neighborhood range. If a pixel with a mask value of 1 exists, it indicates that the current pixel coordinates belong to the potential extended region of microbubble influence, and its mask value is modified from the initial 0 to 1. If it does not exist, the mask value remains unchanged at 0. When traversing and processing pixels at the edge of the image, if the structuring element part exceeds the boundary of the mask image, the boundary zero-filling method is adopted, that is, it is assumed that the mask value of the pixels outside the boundary is 0, to ensure that the dilation operation of the edge region is complete and without abnormalities. After completing the traversal and operation of all pixel coordinates, a new mask image is generated based on the first bubble contamination mask image, which expands the marked region. This image is the expanded bubble contamination mask image. Its marked region includes both the region corresponding to the originally determined microbubble flickering dominant block and the adjacent potentially affected region.

[0067] In an embodiment of the present invention, the specific steps for removing bubble contamination from the image are defined as follows: Construct an initial grayscale image based on the first image; An initial grayscale image is an intermediate image constructed during image processing. Its pixel coordinates are consistent with those of the first image, and it is used to carry the grayscale information after subsequent de-bubble contamination processing pixel by pixel.

[0068] First, the complete pixel dimension information of the first image is obtained, including the total number of rows and columns of pixels and the effective range of pixel coordinates. The two-dimensional coordinate identifier of each pixel in the first image is defined to ensure that the subsequently constructed image can form a precise coordinate mapping relationship with the first image. Then, based on the pixel dimension information, a blank image frame with the same size as the first image is constructed. The total number of rows and columns of pixels in the blank image frame is strictly consistent with the first image to avoid problems such as coordinate misalignment or size mismatch. At the same time, the image channel type of the blank image frame is set to a single-channel grayscale channel to meet the technical requirements for carrying grayscale information. Next, all pixel coordinates in the blank image frame are initialized pixel by pixel in the order from left to right and from top to bottom. Each pixel coordinate is assigned an initial grayscale value that is within the normal range of grayscale image values ​​and does not interfere with subsequent grayscale information updates, ensuring that each pixel position has basic grayscale information storage capabilities. Finally, an intermediate result image is formed with pixel coordinates that perfectly match the first image, channel type that adapts to grayscale information, and the ability to carry grayscale data after subsequent bubble decontamination processing pixel by pixel. This image is the initialized grayscale image.

[0069] When the mask value of the extended bubble contamination mask image at the pixel coordinates is 0, the pixel gray value of the baseline brightness image at the pixel coordinates is used as the pixel gray value of the initial grayscale image at the pixel coordinates. When the mask value of the extended bubble contamination mask image at a certain pixel coordinate is 0, it indicates that the pixel position is not considered to be affected by the microbubble scintillation effect. The image brightness change at this position remains relatively stable under different hydraulic conditions, which can truly reflect the structural state of the inner wall of the oil passage. Therefore, the pixel gray value at the corresponding pixel coordinate of the baseline brightness image is selected as the pixel gray value of the initial gray value image at that pixel position. This allows the initial gray value image to inherit the stable brightness information under multiple conditions at this position, thereby avoiding the interference of occasional brightness fluctuations under a single condition on the results. Through this processing method, the initial gray value image can accurately characterize the true surface brightness distribution of the inner wall of the oil passage in the area unaffected by bubbles, providing reliable basic image data for subsequent cavitation erosion analysis.

[0070] When the mask value of the expanded bubble contamination mask image at the pixel coordinates is 1, the pixel gray value of the initial grayscale image at the pixel coordinates is set to an invalid value. When the mask value of the expanded bubble contamination mask image at a certain pixel coordinate is 1, it indicates that the pixel position has been determined to be directly or indirectly affected by the microbubble scintillation effect. The image brightness change at this position is mainly dominated by the reflection and refraction changes caused by microbubbles in the oil, and can no longer stably reflect the true structural state of the inner wall of the oil passage. Therefore, the pixel gray value at this pixel coordinate in the initial grayscale image is set to an invalid value, so that the pixel position does not participate in feature calculation and state assessment in the subsequent cavitation erosion analysis. This effectively avoids the interference of the brightness abnormal area affected by microbubbles on the analysis of the true metal surface texture, thereby improving the reliability and accuracy of the monitoring and early warning results.

[0071] Invalid values ​​refer to placeholder values ​​in the initial grayscale image used to identify the corresponding pixel position that is not involved in subsequent cavitation erosion analysis.

[0072] Define the initial image as a bubble-removed contamination image.

[0073] A de-bubble contamination image refers to a grayscale image formed after filtering the oil passage inner wall image based on an extended bubble contamination mask image during image processing. This image retains grayscale information representing the stable surface state of the oil passage inner wall at pixel locations not marked as bubble contamination areas, while pixel locations marked as bubble contamination areas are not included in subsequent analysis. This effectively eliminates the interference of microbubble flickering effect on image brightness, enabling the obtained image to more realistically reflect the actual structural state of the oil passage inner wall.

[0074] S5. Monitor and provide early warning of cavitation erosion status of the inner wall of the oil passage by using de-bubbling contamination images.

[0075] In an embodiment of the present invention, the specific steps for monitoring and early warning of cavitation erosion are as follows: Determine the grayscale gradient magnitude of image micro-patch in the de-bubbling contamination image; Sum the magnitudes of all grayscale gradients to obtain the total grayscale gradient. For each image micro-patch in the de-bubble contamination image, the coordinate range of all pixels contained in the micro-patch is first determined, and the grayscale value of each pixel is read one by one within this range. Then, taking each pixel as the center, the strength of the grayscale change at the pixel position is determined by combining the grayscale changes of its adjacent pixels in the horizontal and vertical directions, thereby obtaining the grayscale gradient amplitude corresponding to the pixel. After determining the grayscale gradient amplitude of all pixels in the image micro-patch, the grayscale gradient amplitudes of all pixels belonging to the same image micro-patch are accumulated one by one according to the pixel coordinates to form the grayscale gradient sum corresponding to the image micro-patch. This grayscale gradient sum is used to reflect the cumulative degree of overall brightness change in the image micro-patch, providing basic data for subsequent analysis of the surface undulation characteristics of the image micro-patch.

[0076] Gray-level gradient magnitude refers to the numerical value used to describe the intensity of gray-level changes between adjacent pixels in a de-bubbling contamination image, reflecting the steepness of the brightness change on the inner wall surface of the oil passage; gray-level gradient sum refers to the value obtained by accumulating the gray-level gradient magnitudes corresponding to all pixels within the same image micro-block, used to characterize the cumulative degree of brightness change of the entire micro-block.

[0077] Divide the sum of grayscale gradients by the total number of pixels in the image micro-block to obtain the block roughness value; Block roughness value is a quantitative description used to characterize the surface undulation and unevenness of a local area of ​​the inner wall of an oil passage. This value reflects whether there are micro-pits, protrusions or texture changes on the metal surface in this local area, and its magnitude reflects the roughness or smoothness of the surface structure in this area.

[0078] The block roughness values ​​of all image micro-blocks are averaged to obtain the global roughness mean. The global roughness mean refers to the overall level obtained by comprehensively characterizing the surface undulation of each local area within the overall observation range of the oil passage inner wall. It is used to reflect the average surface state of the oil passage inner wall on a macro scale. This value reflects the smoothness of the overall structure of the oil passage inner wall and provides an overall reference for judging whether there is abnormal roughness or potential erosion in a local area.

[0079] The roughness of the inner wall surface of an oil duct can be understood as a comprehensive manifestation of the spatial undulations of the surface within a certain area. The gray-level gradient reflects the strength of brightness changes between adjacent positions in an image, corresponding to the amplitude of height changes on the inner wall surface of the oil duct at a microscale. When the gray-level gradient changes of all pixels within an image micro-block are summarized and normalized, the resulting block roughness value can reflect the average intensity of surface undulations in that local area. The overall surface state of the inner wall of the oil duct is composed of multiple local areas. Each image micro-block corresponds to local surface features at different locations. Averaging the block roughness values ​​of all image micro-blocks is equivalent to balancing and integrating the intensity of local surface undulations across the entire observation area, thereby obtaining a global roughness mean that can characterize the overall surface roughness level of the inner wall of the oil duct. This global roughness mean reflects the average undulation characteristics of the inner wall of the oil duct at a macroscale, serving as a reference benchmark for evaluating the overall surface state and comparing abnormally rough areas.

[0080] Based on the numerical comparison results of the global roughness mean and the block roughness value, roughness level is marked for each image micro-block to obtain the erosion risk marking result; Roughness level labeling refers to the graded description of the local area corresponding to the micro-block of the image based on the relative relationship between the block roughness value and the global roughness mean, which is used to distinguish different degrees of surface anomaly. Erosion risk labeling results refer to the risk indication information formed for each local area of ​​the inner wall of the oil passage based on the performance of its surface undulation relative to the overall level. This result is used to characterize the probability of material erosion in each local area.

[0081] For each image micro-block obtained from the de-bubbling contamination image, the block roughness value of each image micro-block is first extracted and processed by grayscale gradient accumulation and normalization. A complete roughness value set is constructed by collecting the block roughness values ​​of all analyzable image micro-blocks. This set is then sorted in ascending order of value. The total number of image micro-blocks in the set is counted, and the first and last positions of the sorted set are determined. The sorted set is then evenly divided into three continuous intervals: a low value interval, a middle value interval, and a high value interval. Each interval contains one-third of the total number of image micro-blocks. Simultaneously, the arithmetic mean of all values ​​in the roughness value set is calculated to obtain the global roughness mean, which serves as a reference benchmark for the overall surface roughness level of the oil passage inner wall. Subsequently, each image micro-block is judged separately. First, its block roughness value is located within the sorted set, and its specific sorting position is determined. Then, the difference and ratio between the block roughness value and the global roughness mean are calculated. If the block roughness value is located in the middle value interval, and its roughness value is within the global roughness mean, the roughness is considered acceptable. If the ratio is between 0.8 and 1.2, the surface undulation of the oil duct inner wall region corresponding to the image micro-block is considered to be consistent with the overall average level, with no obvious abnormal roughness features, and it is marked as a low erosion risk state. If the block roughness value is in the high value range, and its ratio with the global roughness average is greater than 1.5, and it is in the top 30% of the sorted set, the surface undulation of the region corresponding to the image micro-block is considered to be significantly higher than the overall level, with obvious abnormal roughness and cavitation erosion features, and it is marked as a high erosion risk state. If the block roughness value is in the low value range, and its ratio with the global roughness average is less than 0.5, the surface of the region is considered to be too smooth but without erosion-related anomalies, and it is also marked as a low erosion risk state. The above interval division, numerical comparison, feature judgment and level marking operations are performed on all image micro-blocks in sequence, and finally a complete marking result set covering the entire oil duct inner wall observation area and containing the erosion risk level of each image micro-block is formed. This set is the erosion risk marking result.

[0082] A cavitation erosion risk distribution map is generated based on the erosion risk labeling results; A cavitation erosion risk distribution map is an image representation formed by displaying the erosion risk marking results according to the spatial position of the inner wall of the oil passage in the image. It is used to intuitively reflect the distribution of cavitation erosion risk at different locations on the inner wall of the oil passage.

[0083] First, the total number of rows and columns of pixels in the de-bubbling contamination image, as well as the effective spatial coordinate range of the oil passage inner wall in the image, are retrieved. Based on this dimensional information, a blank risk mapping image framework is constructed that perfectly matches the spatial dimensions of the de-bubbling contamination image and has a one-to-one correspondence between pixel coordinates. This ensures that the spatial location of the risk distribution accurately matches the actual location of the oil passage inner wall. Then, a quantitative mapping rule is established between the erosion risk marking results and the visualization features. This rule is determined by analyzing visual recognition and distinguishability. Different RGB color channel values ​​are set for different levels in the erosion risk marking results. The color channel values ​​corresponding to different levels must meet the requirement that the grayscale difference between adjacent levels is not less than 50 to ensure clear visual distinction. Next, all image micro-blocks are traversed one by one from left to right and top to bottom, locating the upper left and lower right pixel coordinates of each image micro-block in the blank risk mapping image framework. The risk risk markers for each micro-block are clearly defined, and their corresponding erosion risk markers are extracted. Based on the established quantization mapping rules, the corresponding RGB color values ​​are filled pixel by pixel into the corresponding coordinate area of ​​the risk mapping image frame. After filling all image micro-blocks, a risk level legend is added to the designated edge area of ​​the risk mapping image. The legend clearly marks the RGB color samples corresponding to each erosion risk marker and the corresponding surface condition description. At the same time, an image coordinate system label and an actual size scale are added to accurately locate the actual physical location of the risk area on the inner wall of the oil passage. Finally, the generated image is subjected to contrast enhancement processing to improve the visual distinction between areas of different risk levels, ensuring that high-risk areas stand out more visually. Ultimately, an image is formed that fully reflects the distribution of erosion risk in various local areas of the inner wall of the oil passage in terms of spatial location. This image is the cavitation erosion risk distribution map.

[0084] The cavitation erosion status of the inner wall of the oil passage is monitored and warned by using a cavitation erosion risk distribution map.

[0085] Cavitation erosion refers to the changes in the surface structure of the inner wall of an oil passage due to cavitation during the flow of lubricating oil. It manifests as pitting, pitting, or surface deterioration of local materials. By analyzing the risk distribution map of cavitation erosion, the development trend of cavitation erosion on the inner wall of the oil passage can be judged and warned.

[0086] After generating the cavitation erosion risk distribution map, the pixel dimension information of the de-bubbling contamination image corresponding to the distribution map, the spatial coordinate mapping relationship of the inner wall of the oil passage in the image, and the physical structure parameters of the integrated valve block are retrieved first. A precise correspondence between the pixel coordinates of the distribution map and the actual physical location of the oil passage is established. The specific location, extension range, spatial distribution pattern, and relative position of the low-risk and high-risk erosion states within the oil passage and the key structure of the oil passage are clarified. At the same time, historical normal roughness data of valve blocks of the same type in the plant, factory reference roughness data, past cavitation erosion failure case data, and historical risk distribution data of multiple batches of monitoring are collected to construct a complete risk assessment reference database. By statistically analyzing the area proportion range, distribution characteristics of the low-risk erosion area under normal conditions and the critical judgment parameters of the high-risk erosion area in the reference database, reasonable early warning judgment criteria are determined.

[0087] Subsequently, a comprehensive analysis of the cavitation erosion risk distribution map was conducted. The area, area percentage, continuous distribution length, maximum continuous area size, and distribution density of both low and high cavitation risk states were statistically analyzed. The concentration and diffusion trends of high cavitation risk areas were examined, and whether they overlapped with or were adjacent to historical fault areas. These characteristic parameters were quantitatively compared with the normal parameter ranges in the reference database in multiple dimensions. The deviation of the characteristic parameters of high cavitation risk areas from the normal range was calculated. If the area percentage of high cavitation risk areas exceeded the safe percentage range statistically determined by the reference database, or the maximum continuous area size exceeded the critical value determined based on historical fault data, or if they exhibited a concentrated, contiguous distribution and overlapped with easily eroded key structures such as oil passage bends and borehole intersections, it was determined to be a warning state requiring emergency response. If only a small number of scattered high cavitation risk areas existed, with a small area percentage, no obvious diffusion trend, and deviation within the allowable fluctuation range of the reference database, it was determined to be a warning state requiring routine monitoring. If all areas in the distribution map were low cavitation risk states, and all characteristic parameters were highly consistent with the normal data in the reference database, it was determined to be a state without warning.

[0088] Finally, the complete monitoring and analysis results, including the early warning status determination conclusion, the specific physical location of the risk area, the details of characteristic parameter comparison, and the prediction of risk development trends, are written into the database of the lubricating oil station hydraulic component data monitoring and early warning system. If the early warning status is determined to require emergency response, the system immediately displays a prominent real-time alarm prompt on the central control screen or monitoring interface, clearly marking the oil passage location, risk status, and potential safety impact corresponding to the risk area, automatically generating an emergency maintenance work order, and clearly suggesting the time window for shutdown and maintenance, key inspection parts, and testing methods. If the early warning status is determined to require routine tracking, the system generates a routine maintenance work order, prompting that the relevant areas be subject to key inspection, re-monitoring, and trend tracking during the next scheduled maintenance. If the early warning status is determined to be non-alert, only the monitoring results, characteristic parameters, and risk distribution are recorded for subsequent traceability and long-term trend analysis, while the historical data in the reference database is updated. Through the above complete process, accurate monitoring, scientific determination, and timely early warning of the cavitation erosion state of the oil passage inner wall are achieved, providing a reliable guarantee for the safe and stable operation of the lubricating oil station hydraulic system.

[0089] This invention also provides an intelligent data monitoring and early warning system for hydraulic components in a lubrication oil station, the system comprising: The alignment module is used to construct an aligned image set based on the first and second images of the inner wall of the oil passage; The baseline module is used to construct a baseline brightness image based on an aligned set of images; The flicker module is used to generate the average deviation of the block under all operating conditions based on the baseline brightness image, and to perform flicker determination on each image microblock corresponding to the first image to obtain the microbubble flicker dominant block; The mask module is used to construct a mask and perform morphological dilation processing on the microbubble scintillation dominant block to obtain an extended bubble contamination mask image; a de-bubble contamination image is defined based on the extended bubble contamination mask image; The early warning module is used to monitor and provide early warning of the cavitation erosion state of the inner wall of the oil passage through debubbling contamination images.

[0090] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for monitoring and early warning of hydraulic component data of a lubricating oil station with intelligent processing, characterized in that, Includes the following steps: Construct an aligned image set based on the first and second images of the inner wall of the oil passage; Construct a baseline brightness image based on an aligned image set; The average deviation of the block under all operating conditions is generated based on the baseline brightness image, and the flicker is determined for each microblock corresponding to the first image to obtain the microbubble flicker dominant block; Masking and morphological dilation of the microbubble scintillation dominant block yielded an extended bubble contamination mask image. Debubbling-free images are defined based on extended bubble contamination mask images; The cavitation erosion state of the inner wall of the oil passage is monitored and warned by debubbling contamination images.

2. The method for monitoring and early warning of hydraulic component data of an intelligent processed lubricating oil station according to claim 1, characterized in that, The specific steps for constructing the aligned image set are as follows: Positioning the integrated valve block of the hydraulic system in the lubrication station; With the position of the inner wall of the oil passage of the integrated valve block remaining unchanged, the hydraulic system is sequentially set to different hydraulic conditions; wherein, the different hydraulic conditions include at least: a first hydraulic condition and a second hydraulic condition; Under the first hydraulic condition, acquire the first image of the inner wall of the oil passage; Under the second hydraulic condition, a second image of the inner wall of the oil passage is acquired; Use the first image as the reference image; Determine the spatial transformation relationship between the second image and the reference image; Based on the spatial transformation relationship, the second image is subjected to coordinate transformation to obtain the second aligned image; Construct an alignment image set based on the reference image and the second alignment image.

3. The method of claim 1, wherein the method further comprises: A baseline brightness image is constructed based on an aligned image set, including: Get the grayscale value at pixel coordinates for each aligned image in the aligned image set; The median of the grayscale values ​​is calculated to obtain the baseline grayscale value; A baseline brightness image is constructed based on all baseline grayscale values.

4. The intelligent processed lubricating oil station hydraulic component data monitoring and early warning method according to claim 3, characterized in that, The mean deviation of the entire operating block is generated based on the baseline brightness image, including: Subtract the baseline gray value of the baseline brightness image from the gray value of the aligned image to obtain the brightness deviation value; The observation area of ​​the first image is divided into multiple image micro-blocks; Under each hydraulic condition, the brightness deviation values ​​at all pixel coordinates in the image micro-block are summed to obtain the total brightness deviation. Divide the sum of brightness deviations by the total number of pixels in the image micro-blocks to obtain the average block deviation. The average block deviation of the image micro-block under all hydraulic conditions is calculated to obtain the average block deviation under all conditions.

5. The intelligent processed lubricating oil station hydraulic component data monitoring and early warning method according to claim 4, characterized in that, The specific steps to obtain the microbubble scintillation dominant block are as follows: Based on the numerical comparison between the average block deviation and the average block deviation under all working conditions, the image micro-blocks are marked with status to obtain the marking results. Based on the labeling results of the image micro-blocks under all hydraulic conditions, a sequence of operating conditions for each image micro-block is generated; By analyzing the flickering of each image micro-block using the operating condition sequence, the dominant microbubble flickering block is obtained.

6. The intelligent processed lubricating oil station hydraulic component data monitoring and early warning method according to claim 1, characterized in that, The specific steps to obtain the expanded bubble contamination mask image are as follows: Construct a mask image based on the first image; Set the mask value of all pixel coordinates in the mask image to 0; In the mask image, the mask value corresponding to all pixel coordinates belonging to the microbubble scintillation dominant block is set from 0 to 1 to obtain the first bubble contamination mask image; Morphological dilation processing is performed on the first bubble contamination mask image to obtain an expanded bubble contamination mask image.

7. The intelligent processed lubricating oil station hydraulic component data monitoring and early warning method according to claim 3, characterized in that, The specific steps for defining a de-bubble contamination image are as follows: Construct an initial grayscale image based on the first image; When the mask value of the extended bubble contamination mask image at the pixel coordinates is 0, the pixel gray value of the baseline brightness image at the pixel coordinates is used as the pixel gray value of the initial grayscale image at the pixel coordinates. When the mask value of the expanded bubble contamination mask image at the pixel coordinates is 1, the pixel gray value of the initial grayscale image at the pixel coordinates is set to an invalid value. Define the initial image as a bubble-removed contamination image.

8. The method of claim 1, wherein the method is a method of monitoring and early warning of hydraulic component data of an intelligent processed lubricating oil station. The specific steps for monitoring and early warning of cavitation erosion are as follows: Determine the grayscale gradient magnitude of image micro-patch in the de-bubbling contamination image; Sum the magnitudes of all grayscale gradients to obtain the total grayscale gradient. Divide the sum of grayscale gradients by the total number of pixels in the image micro-block to obtain the block roughness value; The block roughness values ​​of all image micro-blocks are averaged to obtain the global roughness mean. Based on the numerical comparison results of the global roughness mean and the block roughness value, roughness level is marked for each image micro-block to obtain the erosion risk marking result; A cavitation erosion risk distribution map is generated based on the erosion risk labeling results; The cavitation erosion status of the inner wall of the oil passage is monitored and warned by using a cavitation erosion risk distribution map.

9. A smart data monitoring and early warning system for hydraulic components in a lubrication oil station, characterized in that, The system includes: The alignment module is used to construct an aligned image set based on the first and second images of the inner wall of the oil passage; The baseline module is used to construct a baseline brightness image based on an aligned set of images; The flicker module is used to generate the average deviation of the block under all operating conditions based on the baseline brightness image, and to perform flicker determination on each image microblock corresponding to the first image to obtain the microbubble flicker dominant block; The mask module is used to construct a mask and perform morphological dilation processing on the microbubble scintillation dominant block to obtain an extended bubble contamination mask image; a de-bubble contamination image is defined based on the extended bubble contamination mask image; The early warning module is used to monitor and provide early warning of the cavitation erosion state of the inner wall of the oil passage through debubbling contamination images.

Images