Cable aging detection method and system based on machine vision

Abstract

The invention discloses a cable aging detection method and system based on machine vision, and relates to the technical field of cable aging detection, and the method comprises the steps: carrying out the statistics of preprocessing induction characteristics of cable bending induced by the defect degree of each section in a to-be-detected cable section needing to be detected; predicting cable bending induction positions of each section in the to-be-detected cable section to obtain a first induction bending position and a second induction bending position, comparing the first induction bending position with the second induction bending position to extract a first bending position and difference bending positions, and obtaining a second bending position from the induction bending positions which are missed in judgment in the difference bending positions; and combining the first bending position and the second bending position into a predicted bending position, counting the continuity between the positions in the predicted bending position and the overlapping condition between each position and each position in the section defect concentration position, and comprehensively predicting the aging severity of the cable to be detected. According to the cable aging detection method and system based on machine vision provided by the invention, the cable aging detection accuracy can be improved.

Description

Technical Field

[0001] This application relates to the technical field of cable aging detection, and particularly to a cable aging detection method and system based on machine vision. Background Art

[0002] Through image acquisition, processing, and analysis, machine vision technology can accurately identify and locate tiny defects or abnormalities on the surface of cables, avoiding problems such as false detection and missed detection caused by human factors, thereby improving the detection accuracy.

[0003] In traditional technologies, manual detection methods are often used for cable aging detection. It mainly relies on the historical experience of manual detection for discrimination, which has strong subjectivity, resulting in a high probability of misjudgment, and fails to further consider the different correlation effects between different defect sections of the cable, greatly reducing the efficiency and accuracy of the detection work. Summary of the Invention

[0004] To overcome the deficiencies of the above-mentioned prior art, this application provides a cable aging detection method and system based on machine vision.

[0005] In a first aspect, a cable aging detection method based on machine vision provided by this application includes:

[0006] Obtain a cable to be tested for aging detection, divide the cable to be tested into several blocks to obtain cable sections to be tested, and statistically analyze the preprocessing-induced features that induce cable bending in terms of the defect degree of each section in the cable sections to be tested;

[0007] According to the preprocessing-induced features, predict the cable bending-induced positions of each section in the cable sections to be tested to obtain a first induced bending position and a second induced bending position, and compare the first induced bending position and the second induced bending position to extract a first bending position and a differential bending position;

[0008] Obtain a second bending position from the induced bending positions misjudged among the differential bending positions, and combine the first bending position and the second bending position into a predicted bending position;

[0009] Statistically analyze the continuity between each position in the predicted bending positions to obtain a first continuity and a second continuity, and statistically analyze the overlapping situation between each position in the predicted bending positions and each position in the section defect concentration positions to obtain a first superposition degree and a second superposition degree;

[0010] Based on the first continuity, second continuity, first superposition degree, and second superposition degree, comprehensively predict the severity of aging of the cable to be tested to obtain a cable aging degree detection result.

[0011] Preferably, a cable to be tested for aging detection is obtained, the cable to be tested is divided into several blocks to obtain cable sections to be tested, and image information detection is performed on each section in the cable sections to be tested to obtain section image information;

[0012] The cable defect degree and the defect concentration position in each image in the section image information are counted to obtain section defect degree data and section defect concentration positions;

[0013] The section defect degree data and the section defect concentration positions are combined into current defect feature information;

[0014] The distances between each position in the section defect concentration positions are counted to obtain defect distances;

[0015] According to the current defect feature information, the historical defect feature information and the historical bending induction position of the cable section to be tested in the historical period are obtained;

[0016] The induction correlation coefficient of the historical defect feature information inducing the historical bending induction position is extracted to obtain a bending induction correlation coefficient;

[0017] The bending induction correlation coefficient, the defect distance and the current defect feature information are combined into a preprocessed induction feature.

[0018] Preferably, according to the preprocessed induction feature, the position inducing the cable bending in the cable section to be tested is predicted to obtain a first induced bending position;

[0019] The historical cable aging characteristic data of the cable section to be tested is obtained, and an aging simulation prediction model is trained according to the historical cable aging characteristic data;

[0020] The current defect feature information is input into the aging simulation prediction model for testing to obtain a second induced bending position;

[0021] The same induced bending positions in the first induced bending position and the second induced bending position are extracted to obtain a first bending position, and the different induced bending positions in the first induced bending position and the second induced bending position are obtained to obtain different bending positions.

[0022] Preferably, the ratio between the bending degree of each position of the first bending position in the first induced bending position and the corresponding section defect degree data is obtained to obtain a first correlation ratio;

[0023] The ratio between the bending degree of each position of the first bending position in the second induced bending position and the corresponding section defect degree data is obtained to obtain a second correlation ratio;

[0024] The average value of the first correlation ratio and the second correlation ratio is calculated to obtain a pre-used correlation ratio;

[0025] According to the pre-used correlation ratio, the induced bending positions that are missed in the differential bending positions are extracted to obtain the second bending positions;

[0026] The first bending position and the second bending position are combined to form a predicted bending position.

[0027] Preferably, the continuity between each position in the predicted bending position is statistically obtained to obtain a to-be-detected continuity. The to-be-detected continuity is divided into a first continuity and a second continuity. The first continuity refers to the adjacent continuity between the bending positions with an adjacent relationship in the predicted bending position, and the second continuity refers to the interval continuity between the bending positions with a non-adjacent relationship in the predicted bending position;

[0028] The overlapping situation between each position in the predicted bending position and each position in the section defect concentration position is statistically obtained to obtain a to-be-detected superposition degree. The to-be-detected superposition degree is divided into a first superposition degree and a second superposition degree. The first superposition degree refers to the overlapping coverage area of the bending positions that overlap with the section defect concentration position in the predicted bending position, and the second superposition degree refers to the overlapping coverage area of the bending positions that do not overlap with the section defect concentration position in the predicted bending position. The second superposition degree is zero.

[0029] Preferably, the first continuity is respectively combined with the first superposition degree and the second superposition degree to obtain a first preprocessing feature and a second preprocessing feature;

[0030] The second continuity is respectively combined with the first superposition degree and the second superposition degree to obtain a third preprocessing feature and a fourth preprocessing feature;

[0031] The first preprocessing feature, the second preprocessing feature, the third preprocessing feature, and the fourth preprocessing feature are combined to form a condition influence feature set;

[0032] The condition influence feature set, the section defect degree data, and the bending degree of each position of the predicted bending position are input into an aging simulation prediction model to perform a comprehensive test on the aging severity of the to-be-detected cable to obtain a cable aging degree detection result.

[0033] In a second aspect, a cable aging detection system based on machine vision includes:

[0034] A cable section segmentation unit, configured to obtain a to-be-detected cable to be subjected to aging detection, divide the to-be-detected cable into several blocks to obtain to-be-detected cable sections, and statistically obtain preprocessing induced features of the cable bending induced by the defect degree of each section in the to-be-detected cable sections;

[0035] The first bending section prediction unit is configured to predict the cable bending-induced positions of each section in the cable section to be measured based on the preprocessed induced features, obtain the first induced bending position and the second induced bending position, and compare the first induced bending position and the second induced bending position to extract the first bending position and the differential bending position;

[0036] The second bending section prediction unit is configured to obtain the second bending position from the induced bending positions that are misjudged among the differential bending positions, and the first bending position and the second bending position are combined into the predicted bending position;

[0037] The correlation condition statistics unit is configured to statistically calculate the continuity between each position in the predicted bending position to obtain the first continuity and the second continuity, and statistically calculate the overlapping condition between each position in the predicted bending position and each position in the section defect concentration position to obtain the first superposition degree and the second superposition degree;

[0038] The aging comprehensive prediction unit is configured to comprehensively predict the severity of aging of the cable to be measured based on the first continuity, the second continuity, the first superposition degree, and the second superposition degree to obtain the cable aging degree detection result.

[0039] Compared with the prior art, the present invention has the following characteristics and beneficial effects:

[0040] The cable to be tested is segmented into multiple sections to obtain cable sections to be tested. In order to further analyze the correlation between defective parts on the cable, by detecting image defect information, the defect degree data of each section in the cable sections to be tested is preliminarily counted. Considering that there is a varying degree of induction of cable bending between the defect degree data of adjacent sections, the prediction of the bending induction position is carried out under two prediction conditions. One is to predict the induced bending position according to the bending induction correlation coefficient statistically obtained from the correlation coefficient between historical defect feature information and historical bending induction positions. The other is to predict the induced bending position according to the aging simulation prediction model trained with historical cable aging feature data. Then, based on the comparison between the predicted induced bending positions, the accurate predicted bending position is extracted. Through prediction processing using two prediction methods, the accuracy of the predicted bending position is verified to eliminate misjudged prediction results, improving the rigor of the entire detection process. Further, the continuity status between bending positions in the predicted bending positions is statistically analyzed: one is adjacent continuity, and the other is interval continuity. Moreover, the superposition status between the bending position and the defect concentration position in the predicted bending positions is statistically analyzed: one is having an overlapping coverage area, and the other is not having an overlap. By making a differential judgment on these four situations, the influence factors in diverse situations are fully considered, and the cable aging degree is accelerated to different degrees. The correlation influence between diverse information is utilized to improve the reliability and accuracy of the final cable aging degree detection result. Brief Description of the Drawings

[0041] Figure 1 It is a block diagram of the steps of a cable aging detection method based on machine vision mainly embodied in this embodiment.

[0042] Figure 2 It is a block diagram of the structure of a cable aging detection system based on machine vision mainly embodied in this embodiment. Detailed Embodiment

[0043] The present invention will be further described in detail below in conjunction with the following embodiments.

[0044] Refer to Figure 1 , a cable aging detection method based on machine vision, the method includes the following steps:

[0045] S1. Obtain the cable to be tested for aging detection, segment the cable to be tested into several blocks to obtain cable sections to be tested, and statistically analyze the preprocessing induction features of the defect degree of each section in the cable sections to be tested that induce cable bending.

[0046] S2. Predict the cable bending induced positions of each section in the cable section to be measured according to the preprocessing induced features to obtain the first induced bending position and the second induced bending position, and compare the first induced bending position and the second induced bending position to extract the first bending position and the differential bending position.

[0047] S3. Obtain the second bending position from the induced bending positions misjudged among the differential bending positions, and combine the first bending position and the second bending position into the predicted bending position.

[0048] S4. Statistically obtain the first continuity and the second continuity for the continuity between each position in the predicted bending positions, and statistically obtain the first superposition degree and the second superposition degree for the overlapping situation between each position in the predicted bending positions and each position in the section defect concentration positions.

[0049] S5. Comprehensively predict the aging severity of the cable to be measured according to the first continuity, the second continuity, the first superposition degree and the second superposition degree to obtain the cable aging degree detection result.

[0050] Specifically, the cable to be measured is segmented into multiple sections to obtain the cable sections to be measured. In order to further analyze the correlation between the defective parts on the cable, by detecting the image defect information, the defect degree data of each section in the cable sections to be measured are preliminarily statistically obtained. Considering that there are different degrees of induced cable bending between the defect degree data of adjacent sections, the prediction of the bending induced positions is carried out under two prediction conditions. One is to predict the induced bending position according to the correlation coefficient of bending induction statistically obtained between the historical defect feature information and the historical bending induced position. The other is to predict the induced bending position according to the aging simulation prediction model trained by the historical cable aging feature data. Then, according to the comparison between the predicted induced bending positions, the accurate predicted bending position is extracted. Through the prediction processing by two prediction methods, the accuracy verification of the predicted bending position is realized, the misjudged prediction results are eliminated, and the rigor of the whole detection process is improved. Further, the continuity conditions between the bending positions in the predicted bending positions are statistically obtained: one is the adjacent continuity, and the other is the interval continuity. Moreover, the superposition degree conditions between the bending positions and the defect concentration positions in the predicted bending positions are statistically obtained: one is the overlapping coverage area, and the other is the non-overlapping. By making a differential judgment on the four conditions, the influence factors in diverse situations are fully considered, the acceleration of the cable aging degree is affected to different degrees, and the correlation influence between diverse information is utilized to improve the reliability and accuracy of the final cable aging degree detection result.

[0051] Specifically, step S1 includes the following sub-steps:

[0052] Obtain the cable to be tested for aging detection, divide the cable to be tested into several blocks to obtain the cable sections to be tested, and perform image information detection on each section in the cable sections to be tested to obtain the section image information.

[0053] Statistically analyze the cable defect degree and the defect concentration position in each image in the section image information to obtain the section defect degree data and the section defect concentration position.

[0054] The section defect degree data and the section defect concentration position are combined into the current defect feature information.

[0055] Statistically analyze the distance between each position in the section defect concentration position to obtain the defect distance.

[0056] According to the current defect feature information, obtain the historical defect feature information and the historical bending induction position of the cable section to be tested in the historical period.

[0057] Extract the induction correlation coefficient of the historical defect feature information inducing the historical bending induction position to obtain the bending induction correlation coefficient.

[0058] The bending induction correlation coefficient, the defect distance, and the current defect feature information are combined into the preprocessed induction feature.

[0059] Specifically, for the cable section to be measured (divided into section 1, section 2, section 3, and section 4), section image information (i.e., collected from the machine's perspective and processed into grayscale images to more clearly detect the specific defect degree and coverage location status), section defect degree data (defects include scratches, insulation layer damage degree, holes, etc. The defect degree here is a comprehensive evaluation value combined with various damage situations. If they are 2%, 5%, 4%, and 4% respectively), and section defect concentration locations (for example, section 2 is further divided into three location areas: a, b, and c. Among them, a is adjacent to section 1, b is between a and c, and c is adjacent to section 3. If the defect concentration location in section 2 is a, and so on for sections 1, 3, and 4), defect distance (if the defect concentration locations in sections 1, 2, 3, and 4 are c, a, b, and a respectively, then the distances between c - a, a - b, and b - a are statistically calculated), bending-induced correlation coefficient (such as historical bending-induced locations: taking one section as an example, if the induced location in section 1 of the historical period detected in the first historical period is a, and the defect degree in the corresponding historical period in the historical defect feature information is 2% and the concentration location is b, then the induced location in section 1 of the historical period detected in the second historical period is b, and the defect degree in the corresponding historical period in the historical defect feature information is 5% and the concentration location is b, and then the induced location in section 1 of the historical period detected in the third historical period is c, and the defect degree in the corresponding historical period in the historical defect feature information is 3% and the concentration location is b. Then, the estimated bending-induced correlation coefficient from the first historical period to the second historical period is: the position change degree from a to b is 1 / 3, and the defect degree increases by 3%, so the estimated bending-induced correlation coefficient is about 1%. Similarly, if the estimated bending-induced correlation coefficient from the second historical period to the third historical period is about 1.3%, and the average of the two is obtained as 1.2%, that is, the final bending-induced correlation coefficient is about 1.2%).

[0060] Specifically, step S2 includes the following sub-steps:

[0061] According to the preprocessed induced characteristics, predict the cable bending position induced in the cable section to be measured to obtain the first induced bending position.

[0062] Obtain the historical cable aging characteristic data of the cable section to be measured, and train an aging simulation prediction model based on the historical cable aging characteristic data.

[0063] Input the current defect feature information into the aging simulation prediction model for testing to obtain the second induced bending position.

[0064] Extract the same induced bending positions from the first induced bending position and the second induced bending position to obtain the first bending position, and extract the different induced bending positions from the first induced bending position and the second induced bending position to obtain the differential bending position.

[0065] Specifically, for the first induced bending position (for example, [(W - w)+d] / 2 = x / y, where W - w is the interval between the required bending position and the position of concentrated section defects, d is the distance between defects, x is the bending induction correlation coefficient, and y is the section defect degree data, from which the bending position of W can be obtained, and so on for sections 1, 2, 3, and 4 in sequence for prediction. That is, if the first induced bending positions are a, c, a, b respectively), the aging simulation prediction model (i.e., the historical cable aging characteristic data includes the above - mentioned historical cable defect degree record data, bending induction positions, and concentrated defect positions and other characteristic information. The change rules of the historical cable aging characteristic data in different historical periods are plotted as a change path curve graph. For example, the x - axis is the required bending position, and the y - axis includes several change path curve graphs of information: the change path curve graph of historical cable defect degree record data, bending induction positions, and concentrated defect positions and other characteristic information. The corresponding positions of these several pieces of information concentrated on the x - axis are used as the required bending positions, and the trained curve graph is the trained aging simulation prediction model. It also includes the cable comprehensive aging degree values corresponding to the historical cable defect degree record data, bending induction positions, and concentrated defect positions and other characteristic information in the historical cable aging characteristic data. According to the data changes in different historical periods, the corresponding aging degree development curve graph is drawn, that is, according to the change degrees of multiple historical change characteristic data, the cable aging degree development prediction is carried out), the second induced bending position (such as matching the current defect characteristic information with the historical defect characteristic information in the aging degree development curve graph of the trained aging simulation prediction model to correspondingly match the second induced bending position. If it is b, c, a, c), the first bending position (then it is c, a), and the differential bending position (then it is b, c).

[0066] Specifically, step S3 includes the following sub - steps:

[0067] Calculate the ratio between the bending degree of each position of the first bending position in the first induced bending position and the corresponding section defect degree data to obtain the first correlation ratio.

[0068] Calculate the ratio between the bending degree of each position of the first bending position in the second induced bending position and the corresponding section defect degree data to obtain the second correlation ratio.

[0069] Calculate the average value of the first correlation ratio and the second correlation ratio to obtain the pre - used correlation ratio.

[0070] According to the pre-used correlation ratio, the induced bending positions missed in the differential bending positions are extracted to obtain the second bending positions.

[0071] The first bending position and the second bending position are combined into the predicted bending position.

[0072] Specifically, such as the first correlation ratio (if the bending degrees of c and a in the first induced bending position are 2% and 3% respectively, and the corresponding section defect degree data are 5% and 6% respectively, then the first correlation ratios are 0.4 and 0.5 respectively), the second correlation ratio (and so on, if they are 0.5 and 0.6 respectively), the pre-used correlation ratio (i.e., (0.4 + 0.5 + 0.5 + 0.6) / 4 = 0.5), the second bending position (i.e., the remaining a and b in the first induced bending position and the remaining b and c in the second induced bending position are all processed in the same way as the first correlation ratio. If they are 0.3, 0.4, 0.3, and 0.5 respectively, then the remaining c predicted bending position in the corresponding second induced bending position is the missed bending position), the predicted bending position (i.e., the c position in section 2, the a position in section 3, and the c position in section 4).

[0073] Specifically, step S4 includes the following sub-steps:

[0074] The continuity between each position in the predicted bending position is statistically analyzed to obtain the to-be-tested continuity, which is divided into the first continuity and the second continuity. The first continuity refers to the adjacent continuity between the bending positions with an adjacent relationship in the predicted bending position, and the second continuity refers to the interval continuity between the bending positions with a non-adjacent relationship in the predicted bending position.

[0075] The overlapping situation between each position in the predicted bending position and each position in the section defect concentration position is statistically analyzed to obtain the to-be-tested superposition degree, which is divided into the first superposition degree and the second superposition degree. The first superposition degree refers to the overlapping coverage area of the bending positions that overlap with the section defect concentration position in the predicted bending position, and the second superposition degree refers to the overlapping coverage area of the bending positions that do not overlap with the section defect concentration position in the predicted bending position. The second superposition degree is zero.

[0076] Specifically, for the continuity to be measured (position c in section 2, position a in section 3, position c in section 4. For example, the continuity between position c in section 2 and position a in section 3 is 0 because one is at the end position of section 1 and the other is at the start position of section 2 and they are adjacent, which is the adjacent continuity: the first continuity. The continuity between position a in section 3 and position c in section 4 is 1 because one is at the start position of section 3 and the other is at the end position of section 4, which is the spaced continuity: the second continuity), and for the superposition degree to be measured (taking position a in section 3 as an example. If position a in section 3 is also a position with concentrated defect degree, then the two conditional factors are in a superposition state, and the aging acceleration degree of this cable section is the largest. And the covered area of the superposition of the two factors is statistically calculated. If the covered area of the bending position is s1, the covered area of the defect is s2, and the overlapping covered area between s1 and s2 is s3, this is the first superposition degree. If there is no position with concentrated defect degree at position a in section 3, then there is no overlapping covered area between s1 and s2, that is, the overlapping covered area is 0, which is the second superposition degree).

[0077] Specifically, step S5 includes the following sub-steps:

[0078] Combine the first continuity with the first superposition degree and the second superposition degree respectively to obtain preprocessing feature one and preprocessing feature two.

[0079] Combine the second continuity with the first superposition degree and the second superposition degree respectively to obtain preprocessing feature three and preprocessing feature four.

[0080] Preprocessing feature one, preprocessing feature two, preprocessing feature three, and preprocessing feature four are combined into a conditional influence feature set.

[0081] Input the conditional influence feature set, the section defect degree data, and the bending degree of each position for predicting the bending position into the aging simulation prediction model to conduct a comprehensive test on the aging severity of the cable to be measured to obtain the cable aging degree detection result.

[0082] Specifically, such as preprocessing feature one (e.g., the feature of the first continuity and the first superposition degree existing simultaneously at position c in section 2, or position a in the section, or position c in section 4), and preprocessing feature two (or, for example, the feature of the first continuity and the second superposition degree existing simultaneously at position c in section 2, or position a in the section, or position c in section 4), preprocessing feature three and preprocessing feature four (with the same explanations as preprocessing feature one and preprocessing feature two, such as the feature of the second continuity and the first superposition degree existing simultaneously at position c in section 2, or position a in the section, or position c in section 4, or the feature of the second continuity and the second superposition degree existing simultaneously at position c in section 2, or position a in the section, or position c in section 4), the cable aging degree detection result (that is, each section in the cable section to be measured, the condition influence feature set: position c in section 2, position a in section 3, position c in section 4 each contains one of the condition influence features, the section defect degree data: the defect degree data of each of sections 1, 2, 3, and 4, and the predicted bending position: position c in section 2, position a in section 3, position c in section 4 are all input into the aging degree development curve graph trained in the aging simulation prediction model for information matching to estimate the aging degree values of sections 1, 2, 3, and 4 respectively, and then the aging degree values of the four sections are comprehensively estimated, that is, the ratio statistics are performed between the historical aging degree values of each section of the cable in different situations detected in different historical periods and the historical comprehensive section aging degree values, and finally the average value of the ratios in different historical periods is obtained to get the conversion rate, that is, the comprehensive cable aging degree detection result is obtained through integration calculation based on the aging degree values of sections 1, 2, 3, and 4 and the conversion rate).

[0083] A cable aging detection system based on machine vision, by applying a cable aging detection method based on machine vision as described above, includes a cable section segmentation unit, a bending section prediction unit one, a bending section prediction unit two, a correlation condition statistics unit, and an aging comprehensive prediction unit, referring to Figure 2, obtain the cable to be tested that needs to be subjected to aging detection through the cable section segmentation unit, segment the cable to be tested into several blocks to obtain the cable sections to be tested, and count the preprocessing-induced characteristics of cable bending induced by the defect degree of each section in the cable sections to be tested; according to the preprocessing-induced characteristics, the bending section prediction unit 1 predicts the cable bending induced positions of each section in the cable sections to be tested to obtain the first induced bending position and the second induced bending position, and compares the first induced bending position and the second induced bending position to extract the first bending position and the differential bending position; the bending section prediction unit 2 obtains the second bending position from the induced bending positions misjudged in the differential bending positions, and the first bending position and the second bending position are combined into the predicted bending position; the correlation condition statistics unit counts the continuity between each position in the predicted bending positions to obtain the first continuity and the second continuity, and counts the overlapping conditions between each position in the predicted bending positions and each position in the section defect concentration positions to obtain the first superposition degree and the second superposition degree; the aging comprehensive prediction unit comprehensively predicts the aging severity of the cable to be tested according to the first continuity, the second continuity, the first superposition degree and the second superposition degree to obtain the cable aging degree detection result.

[0084] The above are all the preferred embodiments of this application, and the protection scope of this application is not limited accordingly. Therefore, all equivalent changes made according to the structure, shape, and principle of this application should be covered within the protection scope of this application.

Claims

1. A cable aging detection method based on machine vision, characterized in that: The following steps are involved: Obtain a cable to be tested that needs to be subjected to aging detection, divide the cable to be tested into several blocks to obtain cable sections to be tested, and count the pre-processing induced characteristics of cable bending induced by the degree of defects in each section of the cable section to be tested; According to the pre-processing induced characteristics, the cable bending induced position of each section in the cable section to be tested is predicted to obtain a first induced bending position and a second induced bending position, and the first induced bending position and the second induced bending position are compared to extract the first bending position and the difference bending position; A second bending position is obtained from the induced bending position that is missed in the differential bending position, and the first bending position and the second bending position are combined into a predicted bending position; The continuity between each position in the predicted bending position is statistically obtained to obtain a first continuity degree and a second continuity degree, and the overlapping conditions between each position in the predicted bending position and each position in the section defect concentration position are statistically obtained to obtain a first superposition degree and a second superposition degree; According to the first continuity degree, the second continuity degree, the first superposition degree and the second superposition degree, the aging severity of the tested cable is comprehensively predicted to obtain a cable aging degree detection result.

2. A method for metal material quality supervision and evaluation according to claim 1, characterized in that: The steps of obtaining a cable to be tested that needs to be subjected to aging detection, dividing the cable to be tested into several blocks to obtain cable sections to be tested, and counting the pre-processing induced characteristics of the cable bending induced by the defect degree of each section in the cable section to be tested are specifically as follows: Acquire a cable to be tested that needs to be tested for aging, divide the cable to be tested into several blocks to obtain cable sections to be tested, and perform image information detection on each section of the cable sections to be tested to obtain section image information; Counting the cable defect degree and defect concentration position in each image in the segment image information to obtain segment defect degree data and segment defect concentration position; The segment defect degree data and the segment defect concentrated position are combined into current defect characteristic information; The distances between the various locations in the defect concentration location of the section are counted to obtain the defect distances; According to the current defect characteristic information, obtaining historical defect characteristic information and historical bending induced positions of the cable section to be tested in historical periods; Extracting the induced correlation coefficient of the historical bending induced position induced by the historical defect characteristic information to obtain the bending induced correlation coefficient; The bending induced correlation coefficient, the distance between defects and the current defect feature information are combined into a pre-processing induced feature.

3. A method for metal material quality supervision and evaluation according to claim 2, characterized in that: The steps of predicting the cable bending induced position of each section in the cable section to be tested according to the preprocessing induced characteristics to obtain a first induced bending position and a second induced bending position, and comparing the first induced bending position and the second induced bending position to extract the first bending position and the difference bending position are specifically as follows: According to the pre-processing induced characteristics, predicting the induced cable bending position in the cable section to be tested to obtain a first induced bending position; Acquire historical cable aging characteristic data of the cable section to be tested, and train an aging simulation prediction model based on the historical cable aging characteristic data; Inputting the current defect characteristic information into an aging simulation prediction model for testing to obtain a second induced bending position; The same induced bending positions between the first induced bending positions and the second induced bending positions are extracted to obtain first bending positions, and the different induced bending positions between the first induced bending positions and the second induced bending positions are extracted to obtain differential bending positions.

4. A method for metal material quality supervision and assessment according to claim 3, characterized in that: The step of obtaining a second bending position from the induced bending position missed in the differential bending position, and combining the first bending position and the second bending position into a predicted bending position is specifically: Comparing the bending degree of each position of the first bending position in the first induced bending position and the corresponding section defect degree data to obtain a first correlation ratio; Comparing the bending degree of each position of the first bending position in the second induced bending position and the corresponding section defect degree data to obtain a second correlation ratio; averaging the first correlation ratio and the second correlation ratio to obtain a pre-use correlation ratio; According to the pre-used correlation ratio, the induced bending position that is missed in the difference bending position is extracted to obtain a second bending position; The first bending position and the second bending position are combined into a predicted bending position.

5. A method for metal material quality supervision and assessment according to claim 4, characterized in that: The steps of performing statistics on the continuity between each position in the predicted bending position to obtain a first continuity degree and a second continuity degree, and performing statistics on the overlapping conditions between each position in the predicted bending position and each position in the segment defect concentration position to obtain a first superposition degree and a second superposition degree are specifically: The continuity between each position in the predicted bending position is statistically obtained to obtain the continuity to be measured, and the continuity to be measured is divided into a first continuity and a second continuity, the first continuity refers to the adjacent continuity between the bending positions in the predicted bending position, and the second continuity refers to the interval continuity between the bending positions in the predicted bending position that are not adjacent to each other; The overlapping conditions between each position in the predicted bending position and each position in the segment defect concentration position are statistically analyzed to obtain the superposition degree to be measured. The superposition degree to be measured is divided into a first superposition degree and a second superposition degree. The first superposition degree refers to the overlapping coverage area of ​​the bending positions in the predicted bending position that overlap with the segment defect concentration position, and the second superposition degree refers to the overlapping coverage area of ​​the bending positions in the predicted bending position that do not overlap with the segment defect concentration position. The second superposition degree is zero.

6. A method for metal material quality supervision and assessment according to claim 5, characterized in that: The step of comprehensively predicting the aging severity of the cable to be tested to obtain the cable aging degree detection result according to the first continuity degree, the second continuity degree, the first superposition degree and the second superposition degree is specifically as follows: Combining the first degree of continuity with the first degree of superposition and the second degree of superposition respectively to obtain a first preprocessing feature and a second preprocessing feature; Combining the second degree of continuity with the first degree of superposition and the second degree of superposition respectively to obtain preprocessing feature three and preprocessing feature four; The preprocessing feature 1, preprocessing feature 2, preprocessing feature 3 and preprocessing feature 4 are combined into a conditional influence feature set; The conditional influence feature set, the segment defect degree data and the bending degree of each position of the predicted bending position are input into the aging simulation prediction model to conduct a comprehensive test on the aging severity of the cable to be tested to obtain the cable aging degree detection result.

7. A metal material quality supervision and evaluation system, characterized in that: The system is used to implement a metal material quality supervision and evaluation method according to any one of claims 1 to 6, comprising: The cable segmentation unit is used to obtain the cable to be tested that needs to be tested for aging detection, divide the cable to be tested into several blocks to obtain the cable segments to be tested, and count the pre-processing induced characteristics of the cable bending induced by the defect degree of each section in the cable segment to be tested; The bending section prediction unit 1 is used to predict the cable bending induced position of each section in the cable section to be tested according to the preprocessing induced characteristics to obtain the first induced bending position and the second induced bending position, and compare the first induced bending position and the second induced bending position to extract the first bending position and the difference bending position; A second bending section prediction unit is used to obtain a second bending position from an induced bending position that is missed in the difference bending position, and the first bending position and the second bending position are combined into a predicted bending position; An associated condition statistical unit, used for statistically analyzing the continuity between each position in the predicted bending position to obtain a first continuity degree and a second continuity degree, and statistically analyzing the overlapping conditions between each position in the predicted bending position and each position in the section defect concentration position to obtain a first superposition degree and a second superposition degree; The aging comprehensive prediction unit is used to comprehensively predict the aging severity of the cable to be tested according to the first continuity degree, the second continuity degree, the first superposition degree and the second superposition degree to obtain the cable aging degree detection result.

Images