An image analysis method for predicting the life of rigging during use

Abstract

The present invention relates to the technical field of image analysis and processing, and in particular to an image analysis method for predicting the life of rigging during use, the method comprising: collecting rigging images during use, obtaining rigging grayscale images through grayscale processing, calculating the abnormal gradient measurement index of each element in the gradient abnormality monitoring sequence of each pixel point, and calculating the wear edge concave-convex change index of each pixel point, thereby obtaining the rigging wear characteristic index of each pixel point; constructing the local grayscale abnormality index and the rust brightness characteristic change index of each pixel point, based on which the rigging rust characteristic index of each pixel point is obtained; based on the rigging wear characteristic index and the rigging rust characteristic index of each pixel point, the rigging damage characteristic value of each pixel point is obtained, and the rigging life during use is predicted in combination with the rigging damage characteristic value. The present invention improves the accuracy of life prediction of rigging during use.

Description

Technical Field

[0001] The present application relates to the technical field of image analysis and processing, and in particular to an image analysis method for predicting the life of rigging during use. Background Art

[0002] Rigging generally refers to the equipment that matches the cable, such as tensioners, cable clamps, rings and other equipment. Usually, these equipment are collectively referred to as rigging. With the development of the rigging industry, rigging is widely used in many industrial fields, such as port transportation, construction and engineering machinery. As an important part of the cable, rigging often bears the safety of lifting and transportation. The quality of rigging is related to the safety of rigging use, so improving the quality of rigging manufacturing is of great practical significance.

[0003] Usually, during the use of rigging, due to the influence of the use environment, the rigging is prone to wear and corrosion, resulting in low safety in the use of rigging. Usually, when the degree of damage to the rigging during use reaches a certain level, the rigging can no longer be used and the rigging should be replaced in time to ensure the safety of the rigging. Therefore, predicting the service life of the rigging is of great safety significance.

[0004] With the development of image processing technology, most of the images of rigging during use are collected, and the degree of damage of rigging is identified through image feature extraction algorithms, so as to predict the life of rigging during use. Common image feature extraction algorithms include Canny edge detection algorithm, but the edge features extracted by traditional Canny edge detection algorithm are very complex, and the extracted edge features contain normal edges and defective edges. Therefore, the traditional Canny edge detection algorithm cannot accurately identify only defective edges, which affects the assessment of the degree of damage of rigging and makes the accuracy of rigging service life prediction low. Summary of the invention

[0005] In order to solve the above technical problems, the present invention provides an image analysis method for predicting the life of rigging during use, so as to solve the existing problems.

[0006] The image analysis method for predicting the life of rigging during use of the present invention adopts the following technical solution:

[0007] One embodiment of the present invention provides an image analysis method for predicting the life of rigging during use, the method comprising the following steps:

[0008] Collect the rigging images in use, and obtain the rigging grayscale images through grayscale processing;

[0009] According to the gradient amplitude of each pixel point in the rigging grayscale image and the local change of each pixel point, the gradient anomaly detection sequence and edge smoothness change sequence of each pixel point are obtained; according to the difference between the elements in the gradient anomaly detection sequence of each pixel point, the abnormal gradient measurement index of each element in the gradient anomaly monitoring sequence of each pixel point is obtained; according to the difference of elements in the edge smoothness change sequence of each pixel point and the coefficient of variation of all elements, the wear edge concave-convex change index of each pixel point is obtained, and the rigging wear characteristic index of each pixel point is obtained according to the abnormal gradient measurement index and the wear edge concave-convex change index; according to the grayscale value of each pixel point in the rigging grayscale image, the local grayscale variation index of each pixel point is obtained, according to the local grayscale variation index of each pixel point and the brightness value of the pixel point, the rust brightness characteristic variation index of each pixel point is constructed, and according to the rust brightness characteristic variation index of each pixel point and the brightness value of the local pixel point, the rigging rust characteristic index of each pixel point is obtained; based on the rigging wear characteristic index and the rigging rust characteristic index of each pixel point, the rigging damage characteristic value of each pixel point is obtained;

[0010] The prediction result of the rigging life during use is obtained according to the rigging damage characteristic value of each pixel point.

[0011] Preferably, the step of obtaining a gradient anomaly detection sequence and an edge smoothness change sequence of each pixel point according to the gradient amplitude of each pixel point of the rigging grayscale image and the local change of each pixel point comprises:

[0012] The Sobel edge detection algorithm is used to obtain the gradient amplitude of each pixel and all edge pixels in the rigging grayscale image;

[0013] An n×n square window centered on each pixel in the rigging grayscale image is used as the search window for each pixel, and the gradient amplitude of each pixel in the search window of each pixel is used as each element of the gradient anomaly detection sequence of each pixel; a curve composed of all edge pixels in the search window of each pixel is used as the edge curve of the search window, the curvature at the position of each edge pixel on the edge curve is calculated, and the curvature of each edge pixel in the search window of the pixel is used as each element of the edge smoothness change sequence of the pixel.

[0014] Preferably, the step of obtaining the abnormal gradient metric index of each element in the gradient anomaly monitoring sequence of each pixel point according to the difference between the elements in the gradient anomaly detection sequence of each pixel point includes:

[0015] For each element in the gradient anomaly detection sequence of each pixel point, the absolute value of the difference between each element and the element mean of the gradient anomaly detection sequence is calculated, the absolute value of the difference between each element and other elements is obtained, the product of the two absolute values ​​of the difference is calculated, and the average of the products calculated between each element and all other elements is used as the abnormal gradient measurement index of each element.

[0016] Preferably, the step of obtaining the wear edge concave-convex variation index of each pixel point according to the element difference in the edge smoothness variation sequence of each pixel point and the coefficient of variation of all elements includes:

[0017] When the pixel point is an edge pixel point in the rigging grayscale image, the expression of the wear edge concave-convex change index of the pixel point is:

[0018]

[0019] In the formula, C i represents the wear edge concave-convex change index of the i-th pixel in the rigging grayscale image, N represents the number of elements in the edge smoothness change sequence of the i-th pixel in the rigging grayscale image, exp() is an exponential function with a natural constant as the base, and b i,c and b i,c-1 They represent the cth and c-1th elements in the edge smoothness change sequence of the i-th pixel in the rigging grayscale image, respectively. Represents the coefficient of variation of all elements in the edge smoothness change sequence of the i-th pixel in the rigging grayscale image, where ∈ is a value greater than zero;

[0020] On the contrary, the concave-convex variation index of the worn edge of the pixel is 0.

[0021] Preferably, the step of obtaining the rigging wear characteristic index of each pixel point according to the abnormal gradient measurement index and the wear edge concave-convex variation index includes:

[0022] For each pixel point, the mean of the abnormal gradient measurement index of all elements in the gradient anomaly detection sequence of the pixel point is obtained, the information entropy of the abnormal gradient measurement index of all elements in the gradient anomaly detection sequence of the pixel point is calculated, and the sum of the wear edge concave-convex change index of the pixel point and a preset value greater than zero is obtained, and the normalized value of the product of the mean, the information entropy and the sum is used as the rigging wear feature index of the pixel point.

[0023] Preferably, the step of obtaining the local grayscale variation index of each pixel point according to the grayscale value of each pixel point in the rigging grayscale image includes:

[0024] For each pixel in the rigging grayscale image, the mean square error of the grayscale values ​​of all pixels in the search window of the pixel is obtained, the absolute value of the difference between the grayscale value of the pixel and the grayscale values ​​of each pixel in the search window of the pixel is calculated, the mean of the absolute values ​​of the difference calculated between the pixel and all pixels in the search window is obtained, and the product of the mean square error and the mean is used as the local grayscale variation index of the pixel.

[0025] Preferably, the step of constructing the corrosion brightness characteristic variation index of each pixel point according to the local grayscale variation index of each pixel point and the brightness value of the pixel point comprises:

[0026] The rigging image is converted to Lab to obtain the brightness value of each pixel. For each pixel, a clustering algorithm is used to divide the brightness values ​​of all pixels in the search window of the pixel into two clusters. The mean of all brightness values ​​in each cluster is calculated. The cluster with a large mean is used as the rigging feature cluster of the pixel, and the cluster with a small mean is used as the rigging rust feature cluster of the pixel. The rust brightness feature change index F of the i-th pixel is i The expression is:

[0027]

[0028] In the formula, G i represents the number of elements in the rigging corrosion feature cluster of the i-th pixel in the rigging grayscale image, E i Represents the local grayscale variation index of the i-th pixel in the rigging grayscale image, R i , They respectively represent the brightness mean of the rigging feature cluster of the i-th pixel in the rigging grayscale image and the g-th brightness value of the rigging rust feature cluster of the i-th pixel, and ∈ is a preset value greater than zero.

[0029] Preferably, the step of obtaining the rigging corrosion characteristic index of each pixel point according to the corrosion brightness characteristic variation index of each pixel point and the brightness value of the local pixel point includes:

[0030] Calculate the mean of the brightness values ​​of all pixels within the search window of each pixel, and use the mean as the exponent of an exponential function with a natural constant as the base; calculate the ratio between the rigging corrosion feature cluster of each pixel and the number of elements within the rigging feature cluster; and use the normalized result of the product of the calculation result of the exponential function, the ratio and the corrosion brightness feature change index of each pixel as the rigging corrosion feature index of each pixel.

[0031] Preferably, the step of obtaining the rigging damage characteristic value of each pixel point based on the rigging wear characteristic index and the rigging corrosion characteristic index of each pixel point includes:

[0032] The inverse of the product of the rigging wear characteristic index and the rigging corrosion characteristic index of each pixel point is obtained, the inverse is used as the exponent of an exponential function with a natural constant as the base, and the difference between the value 1 and the calculation result of the exponential function is used as the rigging damage characteristic value of each pixel point.

[0033] Preferably, the method of obtaining the prediction result of the rigging life during use according to the rigging damage characteristic value of each pixel point includes:

[0034] The rigging damage feature values ​​of all pixel points constitute a rigging damage feature map, the Otsu threshold is used to obtain the segmentation threshold of the rigging damage feature map, all pixel points whose rigging damage feature values ​​are higher than the segmentation threshold are composed of a rigging damage area, the number of pixels in the rigging damage area is counted, and the ratio between the rigging damage area and the number of pixels in the rigging damage feature map is taken as the rigging damage ratio; the number of days the rigging has been in use is counted, and the result of rounding down the ratio of the number of days to the rigging damage ratio is taken as the total service life of the rigging; the difference between the total service life of the rigging and the number of days is taken as the predicted remaining service life of the rigging.

[0035] The present invention has at least the following beneficial effects:

[0036] The present invention constructs a rigging wear characteristic index based on the analysis of the gradient change and edge smoothness of the rigging grayscale image to reflect the degree of rigging wear in the surrounding area of ​​each pixel in the rigging grayscale image; based on the analysis of the grayscale change characteristics and brightness change characteristics generated when the rigging is rusted, a rigging corrosion characteristic index is constructed to reflect the degree of rust characteristics at different pixel positions in the rigging grayscale image; in order to more accurately evaluate the degree of rigging damage, a rigging damage characteristic value is constructed based on the rigging wear characteristic index and the rigging corrosion characteristic index to reflect the degree of rigging damage represented by each pixel position in the rigging grayscale image. A rigging damage characteristic map is constructed using the rigging damage characteristic value, and the rigging damage ratio is obtained using the rigging damage characteristic map. The prediction result of the rigging life during use is obtained through the rigging damage ratio. The present invention combines the wear characteristics and corrosion characteristics of rigging damage, improves the accuracy of rigging damage degree detection, and improves the accuracy of rigging life prediction during use. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0038] Figure 1 A flowchart of the steps of an image analysis method for predicting the life of rigging during use provided by the present invention. DETAILED DESCRIPTION

[0039] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the image analysis method for life prediction of rigging during use proposed by the present invention, its specific implementation, structure, features and effects, in conjunction with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0040] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0041] The following is a detailed description of a specific scheme of an image analysis method for predicting the life of rigging during use provided by the present invention in conjunction with the accompanying drawings.

[0042] An embodiment of the present invention provides an image analysis method for predicting the life of a rigging during use. Specifically, the following image analysis method for predicting the life of a rigging during use is provided. Figure 1 , the method comprises the following steps:

[0043] Step S001, using a CMOS camera to photograph the rigging in use, obtaining a rigging image, and preprocessing the rigging image to obtain a rigging grayscale image.

[0044] The purpose of this embodiment is to predict the life of the rigging during use, to ensure that the rigging is replaced in time when the degree of damage to the rigging is too large, to avoid serious safety hazards caused by excessive damage to the rigging, and to improve the safety of the use of the rigging. This embodiment uses a CMOS camera to shoot the front of the rigging during use to obtain a rigging image, and the rigging image is an image in the RGB space. In order to improve the accuracy of the subsequent analysis of the degree of damage to the rigging and avoid the influence of noise on the captured image, the rigging image is used as the input of the non-local mean filtering algorithm, the output of the non-local mean filtering algorithm is used as the filtered rigging image, the filtered rigging image is used as the input of the average grayscale algorithm, and the output of the average grayscale algorithm is used as the rigging grayscale image, wherein the average grayscale algorithm and the non-local mean filtering algorithm are both well-known technologies, and the specific process will not be repeated.

[0045] At this point, the preprocessed grayscale image of the rigging is obtained, which is used for subsequent analysis of the degree of damage to the rigging.

[0046] Step S002, construct an abnormal gradient measurement index and a wear edge concave-convex variation index, and then calculate the rigging wear characteristic index; obtain the local grayscale variation index based on the rigging grayscale image, calculate the rust brightness characteristic variation index, and obtain the rigging corrosion characteristic index; obtain the rigging damage characteristic value based on the rigging wear characteristic index and the rigging corrosion characteristic index, and construct a rigging damage characteristic map.

[0047] Rigging is generally used for lifting cargo. Due to the friction of rigging during lifting cargo, it often causes a certain degree of wear on the rigging. As the use cycle of rigging increases, the degree of wear of rigging increases. In addition, due to changes in the use environment, such as changes in air dryness and humidity during use, rigging is prone to rust. The wear and rust of rigging causes great damage to the rigging, which will seriously reduce the service life of the rigging. Therefore, in order to accurately predict the service life of rigging, it is necessary to further analyze the degree of damage of rigging based on images.

[0048] Generally speaking, when the rigging is worn, the worn edge will be different from the normal edge. Wear will cause the edge to be concave, making the edge less smooth. At the same time, the worn edge will show a greater abnormality than the normal edge. Generally, the greater the wear on the edge of the rigging, the greater the damage to the rigging, and the less suitable the rigging is for continued use.

[0049] Specifically, the rigging grayscale image is used as the input of the Sobel edge detection operator, and the output of the Sobel edge detection operator is used as all edge pixels in the rigging grayscale image and the gradient amplitude of each pixel, wherein the Sobel edge detection operator is a well-known technology, and the specific process is not repeated. In order to analyze the degree of wear at different positions in the rigging grayscale image, a square window of size n×n centered on each pixel in the rigging grayscale image is used as the search window for each pixel, and the sequence composed of the gradient amplitudes of all pixels in the search window of each pixel is used as the gradient anomaly detection sequence for each pixel, wherein n is 15 in this embodiment, and can be set by the implementer. For each pixel in the rigging grayscale image, the curve composed of all edge pixels in the search window of each pixel is used as the edge curve of the search window, and the curvature at each edge pixel position on the edge curve of the search window of the pixel is calculated, and the sequence composed of all the curvatures in the search window is used as the edge smoothness change sequence of the pixel, wherein the calculation of the curvature is a well-known technology, and the specific process is not repeated.

[0050] Based on the above analysis, the rigging wear characteristic index of each pixel point in the rigging grayscale image is calculated to reflect the degree of rigging wear in the surrounding area of ​​each pixel point in the rigging grayscale image. The calculation formula is as follows:

[0051]

[0052]

[0053] D i =Norm[X(A i )×P(A i )×(C i +α)]

[0054] In the formula, represents the abnormal gradient metric index of the jth element in the gradient anomaly detection sequence of the i-th pixel in the rigging grayscale image, M represents the number of elements in the gradient anomaly detection sequence of the i-th pixel in the rigging grayscale image, and They represent the jth and sth element values ​​in the gradient anomaly detection sequence of the i-th pixel in the rigging grayscale image, respectively. represents the element mean in the gradient anomaly detection sequence of the i-th pixel in the rigging grayscale image;

[0055] C i represents the wear edge concave-convex change index of the i-th pixel in the rigging grayscale image, i represents the i-th pixel in the rigging grayscale image, Q represents the set of all edge pixels in the rigging grayscale image, N represents the number of elements in the edge smoothness change sequence of the i-th pixel in the rigging grayscale image, exp() is an exponential function with a natural constant as the base, and b i,c and b i,c-1 They represent the cth and c-1th elements in the edge smoothness change sequence of the i-th pixel in the rigging grayscale image, respectively. The coefficient of variation of all elements in the edge smoothness variation sequence of the i-th pixel in the rigging grayscale image, ∈ is a preset value greater than zero to avoid the denominator being 0. In this embodiment, the value is 1, wherein the calculation of the coefficient of variation is a well-known technique, and the specific process is not repeated here;

[0056] D i represents the rigging wear characteristic index of the i-th pixel in the rigging grayscale image, Norm() is the normalization function, X(A i ) represents the mean value of the abnormal gradient metric index of all elements in the gradient anomaly detection sequence of the i-th pixel in the rigging grayscale image, P(A i ) represents the information entropy of the abnormal gradient measurement index of all elements in the gradient abnormal detection sequence of the i-th pixel in the rigging grayscale image, α is a preset value greater than zero to avoid the calculation of the rigging wear characteristic index to be 0, and the value in this embodiment is 0.1, which can be set by the implementer. The calculation of information entropy is a well-known technology, and the specific process will not be repeated here.

[0057] The difference between the element values The larger the value, the greater the difference between the element value and the element mean. The larger the value is, the greater the abnormal gradient change of the surrounding pixels. Since the worn edge will show a larger abnormality than the normal edge, the more likely it is to be in the worn area, the greater the abnormal gradient metric index is.

[0058] For pixels that do not belong to the edge pixel set, they do not have the concave-convex change that reflects the worn edge, and the worn edge concave-convex change index is 0; for pixels that belong to the edge pixel set, when the difference between adjacent elements in the sequence |b i,c -b i,c-1 The larger the value of |, the greater the growth change of the curvature element in the edge smoothness sequence, which to a certain extent reflects the greater degree of concave and convex changes in the edge line in the search window. At the same time, the coefficient of variation The larger it is, the greater the discrete degree of curvature in the edge smoothness sequence is, that is, the worse the smoothness of the edge line in the search window is, the more likely its edge is to be worn, resulting in concave-convex changes in the smooth edge line, and the greater the concave-convex change index of the worn edge;

[0059] The mean value of the abnormal gradient metric index X(A i ) is larger, and the information entropy P(A i ) is larger, the greater the average abnormality of the gradient in the search window, and the greater the chaotic degree of the gradient in the search window. The wear of the local area of ​​the rigging will cause abnormal changes in the grayscale gradient and chaotic gradient changes. It is more likely to be in the area of ​​rigging wear. At the same time, the wear edge concave-convex change index C i The larger it is, the more prominent the features of the worn edge are. The more likely it is to be in the rigging wear area, the greater the rigging wear feature index is.

[0060] Furthermore, due to the influence of the rigging use environment, the surface of the rigging is prone to rust during use. Usually, the rusted area on the rigging will form a large grayscale difference from the normal area, and due to the damage of the coating, the surface roughness of the rusted area of ​​the rigging is higher and the brightness is lower, while the surface of the normal area is relatively smooth, with lower roughness and higher brightness. In order to more accurately predict the service life of the rigging by the degree of damage to the rigging, it is necessary to further analyze the characteristics of the rust on the rigging.

[0061] Specifically, the rigging image in the RGB space is used as the input of the Lab image conversion technology, and the output of the Lab image conversion technology is used as the rigging perception image in the Lab color space. The brightness value of each pixel in the rigging grayscale image is extracted through the brightness channel of the rigging perception image. The Lab image conversion technology is a well-known technology and the specific process will not be elaborated herein.

[0062] For each pixel in the rigging grayscale image, a set of brightness values ​​of all pixels in the search window of the pixel is used as the input of the k-means clustering algorithm, the preset number of clusters is 2, and the k-means clustering algorithm outputs the clustering result of the brightness values ​​of all pixels in the search window of each pixel in the rigging grayscale image, and the clustering result of the brightness values ​​of all pixels in the search window of each pixel contains two clustering clusters; further, the mean of all brightness values ​​in each clustering cluster is calculated, and the clustering cluster with a large mean is used as the rigging feature clustering cluster of the pixel point, and the clustering cluster with a small mean is used as the rigging rust feature clustering cluster of the pixel point, wherein the k-means clustering algorithm is a well-known technology, and the specific process will not be repeated here.

[0063] Based on the above analysis, the rigging corrosion feature index of each pixel in the rigging grayscale image is calculated to reflect the degree of corrosion features at different pixel positions in the rigging grayscale image. The calculation formula is as follows:

[0064]

[0065]

[0066]

[0067] In the formula, E i represents the local grayscale variation index of the i-th pixel in the rigging grayscale image, σ i (f) represents the mean square error of the grayscale values ​​of all pixels in the search window of the i-th pixel in the rigging grayscale image, B represents the number of neighboring pixels in the search window of the i-th pixel in the rigging grayscale image, and f i Represents the gray value of the i-th pixel in the rigging gray image, represents the gray value of the qth neighboring pixel in the search window of the ith pixel in the rigging gray image;

[0068] F i G represents the rust brightness feature change index of the i-th pixel in the rigging grayscale image. i represents the number of elements in the rigging corrosion feature cluster of the i-th pixel in the rigging grayscale image, R i Represents the brightness mean of the rigging feature cluster of the i-th pixel in the rigging grayscale image, represents the g-th brightness value in the rigging corrosion feature cluster of the i-th pixel in the rigging grayscale image, where ∈ is a preset value greater than zero to avoid the denominator being 0. In this embodiment, the value is 1;

[0069] H i represents the rigging corrosion characteristic index of the i-th pixel in the rigging grayscale image, Norm[] is the normalization function, τ i It represents the ratio between the number of elements in the rigging corrosion feature cluster and the rigging feature cluster at the i-th pixel in the rigging grayscale image. exp() is an exponential function with a natural constant as the base. Represents the mean brightness of all pixels in the search window of the i-th pixel in the rigging grayscale image.

[0070] Mean square error σ i (f) is larger, and the grayscale difference between pixels is The larger the value is, the greater the discreteness of the grayscale in the search window is. Since the grayscale difference between the rusted area and the normal area is relatively large, the more likely it is to be in the rusted edge area, and the larger the local grayscale variation index is.

[0071] Local grayscale variation index E i The larger the value, the more likely the pixel is to be in the rust edge area, and the rigging rust will cause the brightness of the rigging surface to decrease, that is, the rust brightness feature of the pixel is more obvious, and the difference between the brightness mean value within the rigging feature cluster and the brightness value within the rigging rust feature cluster is The larger the value, the greater the value of the element in the cluster of the rigging corrosion feature cluster. The smaller it is, the greater the change in brightness in the search window. The smaller the brightness of the pixels represented by the rigging corrosion feature clusters, the more it indicates that corrosion occurs in the search window, and the greater the change index of the corrosion brightness feature.

[0072] The larger the rust brightness characteristic change index is, the more likely it is that rust has occurred on the rigging surface. At the same time, the ratio of the number of elements τ i The larger the brightness, the The smaller it is, to a certain extent, it means that the proportion of rusted pixels in the search window is larger, and it is more likely to be in the area where the rigging is corroded. At this time, the more prominent the characteristics of rigging corrosion are, the larger the rigging corrosion feature index is.

[0073] At this point, the rigging wear characteristic index and the rigging corrosion characteristic index of each pixel point are obtained.

[0074] Based on the above analysis, by extracting the wear characteristics and corrosion characteristics of the rigging surface damage, a rigging damage feature map is constructed based on the wear characteristics and corrosion characteristics of the rigging to more accurately evaluate the degree of rigging damage.

[0075] Furthermore, based on the rigging wear characteristic index and the rigging corrosion characteristic index, the rigging damage characteristic value of each pixel point in the rigging grayscale image is calculated to reflect the degree of rigging damage represented by each pixel point position in the rigging grayscale image. The calculation formula is as follows:

[0076] V i =1-exp(-(D i ×H i ))

[0077] Where V i represents the rigging damage feature value of the i-th pixel in the rigging grayscale image, exp() is an exponential function with a natural constant as the base, D i represents the rigging wear characteristic index of the i-th pixel in the rigging grayscale image, H i Represents the rigging corrosion feature index of the i-th pixel in the rigging grayscale image.

[0078] Rigging wear characteristic index D i The larger the rigging corrosion characteristic index H i The larger it is, the greater the wear characteristics of the local position on the rigging, the greater the rust characteristics, the greater the degree of damage to the rigging at the pixel point i, and the greater the rigging damage characteristic value.

[0079] Furthermore, the grayscale value of each pixel in the rigging grayscale image is replaced with the corresponding rigging damage feature value, and all pixels in the rigging grayscale image are traversed for replacement. The result after replacing all pixels in the rigging grayscale image is recorded as a rigging damage feature map.

[0080] At this point, the rigging damage characteristic diagram is obtained, which is used to predict the service life of the rigging.

[0081] Step S003, obtaining the rigging damage ratio according to the rigging damage characteristic diagram, obtaining the prediction result of the rigging life during use according to the rigging damage ratio, and completing the image analysis process of predicting the life of the rigging during use.

[0082] Furthermore, in order to predict the life of the rigging more accurately, the rigging damage feature map is used as the input of the Otsu threshold segmentation algorithm, the area above the segmentation threshold is used as the rigging damage area, and the output of the Otsu threshold segmentation algorithm is used as the rigging damage area image of the rigging damage feature map, wherein the Otsu threshold segmentation algorithm is a well-known technology and the specific process will not be repeated here.

[0083] The service life of the rigging is predicted based on the image segmentation results. The number of pixels in the rigging damage area in the rigging damage feature map is counted, and the ratio of the number of pixels in the rigging damage area to the number of pixels in the rigging damage feature map is taken as the rigging damage ratio.

[0084] Furthermore, the number of days the rigging has been in use is obtained through the service life of the rigging, and the result of rounding down the ratio of the number of days the rigging has been in use to the damage ratio of the rigging is taken as the total service life of the rigging, and the difference between the total service life of the rigging and the number of days the rigging has been in use is taken as the remaining service life of the rigging, and the remaining service life of the rigging is taken as the prediction result of the rigging life during use.

[0085] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The above is a description of a specific embodiment of this specification. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0086] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0087] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Modifications to the technical solutions recorded in the aforementioned embodiments, or equivalent replacement of some of the technical features therein, do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. An image analysis method for predicting the life of rigging during use, characterized in that: The method comprises the following steps: Collect the rigging images in use, and obtain the rigging grayscale images through grayscale processing; According to the gradient amplitude of each pixel point in the rigging grayscale image and the local change of each pixel point, the gradient anomaly detection sequence and edge smoothness change sequence of each pixel point are obtained; according to the difference between the elements in the gradient anomaly detection sequence of each pixel point, the abnormal gradient measurement index of each element in the gradient anomaly monitoring sequence of each pixel point is obtained; according to the difference of elements in the edge smoothness change sequence of each pixel point and the coefficient of variation of all elements, the wear edge concave-convex change index of each pixel point is obtained, and the rigging wear characteristic index of each pixel point is obtained according to the abnormal gradient measurement index and the wear edge concave-convex change index; according to the grayscale value of each pixel point in the rigging grayscale image, the local grayscale variation index of each pixel point is obtained, according to the local grayscale variation index of each pixel point and the brightness value of the pixel point, the rust brightness characteristic variation index of each pixel point is constructed, and according to the rust brightness characteristic variation index of each pixel point and the brightness value of the local pixel point, the rigging rust characteristic index of each pixel point is obtained; based on the rigging wear characteristic index and the rigging rust characteristic index of each pixel point, the rigging damage characteristic value of each pixel point is obtained; The prediction result of the rigging life during use is obtained according to the rigging damage characteristic value of each pixel point; The step of obtaining the rigging wear characteristic index of each pixel point according to the abnormal gradient measurement index and the wear edge concave-convex change index includes: For each pixel point, the mean of the abnormal gradient measurement index of all elements in the gradient anomaly detection sequence of the pixel point is obtained, the information entropy of the abnormal gradient measurement index of all elements in the gradient anomaly detection sequence of the pixel point is calculated, and the sum of the wear edge concave-convex change index of the pixel point and a preset value greater than zero is obtained, and the normalized value of the product of the mean, the information entropy and the sum is used as the rigging wear feature index of the pixel point.

2. The image analysis method for predicting the life of rigging during use according to claim 1, characterized in that: The method of obtaining a gradient anomaly detection sequence and an edge smoothness change sequence of each pixel point according to the gradient amplitude of each pixel point in the rigging grayscale image and the local change of each pixel point includes: The Sobel edge detection algorithm is used to obtain the gradient amplitude of each pixel and all edge pixels in the rigging grayscale image; A square window of n×n centered on each pixel in the rigging grayscale image is used as a search window for each pixel, and the gradient amplitude of each pixel in the search window of each pixel is used as each element of the gradient anomaly detection sequence of each pixel; The curve composed of all edge pixels in the search window of each pixel point is used as the edge curve of the search window, the curvature at the position of each edge pixel point on the edge curve is calculated, and the curvature of each edge pixel point in the search window of the pixel point is used as each element of the edge smoothness change sequence of the pixel point.

3. The image analysis method for predicting the life of rigging during use according to claim 2, characterized in that: The method of obtaining the abnormal gradient measurement index of each element in the gradient abnormality monitoring sequence of each pixel point according to the difference between the elements in the gradient abnormality detection sequence of each pixel point includes: For each element in the gradient anomaly detection sequence of each pixel point, the absolute value of the difference between each element and the element mean of the gradient anomaly detection sequence is calculated, the absolute value of the difference between each element and other elements is obtained, the product of the two absolute values ​​of the difference is calculated, and the average of the products calculated between each element and all other elements is used as the abnormal gradient measurement index of each element.

4. The image analysis method for predicting the life of rigging during use according to claim 3, characterized in that: The method of obtaining the wear edge concave-convex change index of each pixel point according to the element difference in the edge smoothness change sequence of each pixel point and the coefficient of variation of all elements includes: When the pixel point is an edge pixel point in the rigging grayscale image, the expression of the wear edge concave-convex change index of the pixel point is: In the formula, C i represents the wear edge concave-convex change index of the i-th pixel in the rigging grayscale image, N represents the number of elements in the edge smoothness change sequence of the i-th pixel in the rigging grayscale image, exp() is an exponential function with a natural constant as the base, and b i,c and b i,c-1 They represent the cth and c-1th elements in the edge smoothness change sequence of the i-th pixel in the rigging grayscale image, respectively. Represents the coefficient of variation of all elements in the edge smoothness change sequence of the i-th pixel in the rigging grayscale image, where ∈ is a value greater than zero; On the contrary, the concave-convex variation index of the worn edge of the pixel is 0.

5. The image analysis method for predicting the life of rigging during use according to claim 1, characterized in that: The step of obtaining the local grayscale variation index of each pixel point according to the grayscale value of each pixel point in the rigging grayscale image includes: For each pixel in the rigging grayscale image, the mean square error of the grayscale values ​​of all pixels in the search window of the pixel is obtained, the absolute value of the difference between the grayscale value of the pixel and the grayscale values ​​of each pixel in the search window of the pixel is calculated, the mean of the absolute values ​​of the difference calculated between the pixel and all pixels in the search window is obtained, and the product of the mean square error and the mean is used as the local grayscale variation index of the pixel.

6. The image analysis method for predicting the life of rigging during use according to claim 2, characterized in that: The method of constructing the rust brightness characteristic variation index of each pixel point according to the local grayscale variation index of each pixel point and the brightness value of the pixel point includes: The rigging image is converted to Lab to obtain the brightness value of each pixel. For each pixel, a clustering algorithm is used to divide the brightness values ​​of all pixels in the search window of the pixel into two clusters. The mean of all brightness values ​​in each cluster is calculated. The cluster with a large mean is used as the rigging feature cluster of the pixel, and the cluster with a small mean is used as the rigging rust feature cluster of the pixel. The rust brightness feature change index F of the i-th pixel is i The expression is: In the formula, G i represents the number of elements in the rigging corrosion feature cluster of the i-th pixel in the rigging grayscale image, E i Represents the local grayscale variation index of the i-th pixel in the rigging grayscale image, R i , They respectively represent the brightness mean of the rigging feature cluster of the i-th pixel in the rigging grayscale image and the g-th brightness value of the rigging rust feature cluster of the i-th pixel, and ∈ is a preset value greater than zero.

7. The image analysis method for predicting the life of rigging during use according to claim 6, characterized in that: The method of obtaining the rigging corrosion characteristic index of each pixel point according to the corrosion brightness characteristic change index of each pixel point and the brightness value of the local pixel point includes: Calculate the mean of the brightness values ​​of all pixels within the search window of each pixel, and use the mean as the exponent of an exponential function with a natural constant as the base; calculate the ratio between the rigging corrosion feature cluster of each pixel and the number of elements within the rigging feature cluster; and use the normalized result of the product of the calculation result of the exponential function, the ratio and the corrosion brightness feature change index of each pixel as the rigging corrosion feature index of each pixel.

8. The image analysis method for predicting the life of rigging during use according to claim 1, characterized in that: The method of obtaining the rigging damage characteristic value of each pixel point based on the rigging wear characteristic index and the rigging corrosion characteristic index of each pixel point includes: The inverse of the product of the rigging wear characteristic index and the rigging corrosion characteristic index of each pixel point is obtained, the inverse is used as the exponent of an exponential function with a natural constant as the base, and the difference between the value 1 and the calculation result of the exponential function is used as the rigging damage characteristic value of each pixel point.

9. The image analysis method for predicting the life of rigging during use according to claim 1, characterized in that: The method of obtaining the prediction result of the rigging life during use according to the rigging damage characteristic value of each pixel point includes: The rigging damage feature values ​​of all pixel points constitute a rigging damage feature map, the Otsu threshold is used to obtain the segmentation threshold of the rigging damage feature map, all pixel points whose rigging damage feature values ​​are higher than the segmentation threshold are composed of a rigging damage area, the number of pixels in the rigging damage area is counted, and the ratio between the rigging damage area and the number of pixels in the rigging damage feature map is taken as the rigging damage ratio; the number of days the rigging has been in use is counted, and the result of rounding down the ratio of the number of days to the rigging damage ratio is taken as the total service life of the rigging; the difference between the total service life of the rigging and the number of days is taken as the predicted remaining service life of the rigging.

Images