Flaw detection method and system for alloy steel wire rope production

Abstract

The invention discloses a flaw detection method and system for alloy steel wire rope production, and relates to the technical field of steel wire rope flaw detection, and the flaw detection method comprises the following steps: heating and cooling an alloy steel wire rope, and collecting a thermal imaging image of the alloy steel wire rope to obtain first thermal imaging sample data; performing image cutting processing and image feature processing on the first thermal imaging sample data to obtain second thermal imaging sample data; constructing an original detection model, and training by using the second thermal imaging sample data to obtain an alloy steel wire rope detection model; performing flaw detection on the alloy steel wire rope based on the alloy steel wire rope detection model; the method is used for solving the problems that in the existing steel wire rope flaw detection technology, when flaw detection is carried out through a thermal imaging image of a steel wire rope, only a static thermal imaging image is relied on, dynamic characteristic information existing in the temperature change process of the steel wire rope is ignored, and some tiny defects of the steel wire rope are difficult to find.

Description

Technical Field

[0001] The invention relates to the technical field of steel wire rope flaw detection, and in particular to a flaw detection method and system for alloy steel wire rope production. Background Art

[0002] Wire rope flaw detection technology is a non-destructive testing technology used to detect internal and external damage of wire ropes. Because wire ropes play a key role in the safe operation of many fields, their quality is directly related to the safety of personnel and the normal operation of equipment. Therefore, flaw detection after the production of wire ropes is very important. Comprehensive flaw detection can ensure that the wire ropes meet quality standards.

[0003] Existing flaw detection technologies for wire ropes include electromagnetic flaw detection, ultrasonic flaw detection, machine vision flaw detection, radiographic flaw detection, and thermal imaging flaw detection. In thermal imaging flaw detection, existing technologies often only collect a single or a few static thermal imaging images. Relying on a single or a few static thermal imaging images, it is difficult to capture the dynamic information of internal defects of the wire rope during the heat conduction process; for example, some tiny internal cracks may be difficult to be found in static images due to the relatively stable temperature distribution; and ordinary single or a few static thermal imaging images may also be due to various factors, such as environmental images and the material properties of the wire rope itself, resulting in insufficient defect characteristics. , which leads to low accuracy of flaw detection; for example, in the patent application with publication number CN111678979A, a method and device for detecting defects in wire ropes by eddy current thermal imaging is disclosed. This solution is to perform flaw detection by collecting static thermal imaging images of alloy steel wire ropes. Only defects on the surface of the wire rope can be detected, and defects inside the wire rope are difficult to find; therefore, the existing wire rope flaw detection technology only relies on static thermal imaging images when performing flaw detection through thermal imaging images of wire ropes, and ignores the dynamic characteristic information existing in the temperature change process of the wire rope, making it difficult to find some tiny defects of the wire rope. Summary of the invention

[0004] The present invention aims to solve one of the technical problems in the prior art to at least a certain extent, by heating and cooling the alloy steel wire rope, collecting the thermal imaging image of the alloy steel wire rope, and performing image cropping and image feature processing to obtain a second thermal imaging sample data; constructing an original detection model, and using the second thermal imaging sample data for training to obtain an alloy steel wire rope detection model; and then using the collected thermal imaging image set of the alloy steel wire rope to perform flaw detection on the alloy steel wire rope; so as to solve the problem that the existing steel wire rope flaw detection technology, when performing flaw detection through the thermal imaging image of the steel wire rope, only relies on static thermal imaging images, but ignores the dynamic feature information existing in the temperature change process of the steel wire rope, and some tiny defects of the steel wire rope are difficult to be discovered.

[0005] To achieve the above objectives, in a first aspect, the present application provides a flaw detection method for alloy steel wire rope production, comprising the following steps: Performing heating and cooling treatment on the alloy steel wire rope, and collecting a thermal imaging image of the alloy steel wire rope to obtain first thermal imaging sample data; Performing image cropping and image feature processing on the first thermal imaging sample data to obtain second thermal imaging sample data; An original detection model is constructed and trained using the second thermal imaging sample data to obtain an alloy steel wire rope detection model; Based on the alloy steel wire rope detection model, the alloy steel wire rope is subjected to flaw detection using the collected thermal imaging image set of the alloy steel wire rope.

[0006] Furthermore, heating and cooling the alloy steel wire rope, and collecting a thermal imaging image of the alloy steel wire rope to obtain first thermal imaging sample data include the following sub-steps: A certain number of alloy steel wire ropes are obtained, including defective alloy steel wire ropes and non-defective alloy steel wire ropes, and the obtained alloy steel wire ropes are divided into groups according to whether there are defects and the types of defects, including: a non-defect group, a surface defect group, a broken wire defect group, and an internal defect group; Heating and cooling treatment: Set the heating temperature to T0, and use the heating device to heat the alloy steel wire ropes of the defect-free group, surface defect group, broken wire defect group and internal defect group respectively, and place the heated alloy steel wire ropes in a unified standard environment for natural cooling.

[0007] Furthermore, heating and cooling the alloy steel wire rope, and collecting a thermal imaging image of the alloy steel wire rope to obtain first thermal imaging sample data further includes the following sub-steps: During the natural cooling process, the time window is set to t1, and the grayscale thermal imaging image of the alloy steel wire rope is collected at the first time interval within the time window, and the collection time and the number of collections of each collection are recorded, and the images are classified and stored according to the groups corresponding to the alloy steel wire ropes, marked as the first thermal imaging sample data, and the first time interval is t2; Collecting the grayscale thermal imaging image of the alloy steel wire rope includes: keeping the alloy steel wire rope unbent, placing the alloy steel wire rope parallel to the lens plane of the thermal imaging device, collecting the first grayscale thermal imaging image, and recording the rotation angle of the alloy steel wire rope at this time as the initial angle of 0°, keeping the alloy steel wire rope and the thermal imaging device stationary, rotating the alloy steel wire rope 120° clockwise with the rope direction as the central axis, collecting the second grayscale thermal imaging image, and then rotating the alloy steel wire rope 120° clockwise with the rope direction as the central axis, collecting the third grayscale thermal imaging image, marking the collected grayscale thermal imaging images according to the number of collections and the collection order, and merging and storing multiple grayscale thermal imaging images collected at the same rotation angle.

[0008] Furthermore, performing image cropping and image feature processing on the first thermal imaging sample data to obtain the second thermal imaging sample data includes the following sub-steps: Perform image cropping processing on all grayscale thermal imaging images in the first thermal imaging sample data, and obtain thermal imaging subject sample data after the processing is completed; The image cropping process includes: setting the grayscale threshold to H0, traversing all pixels in the grayscale thermal imaging image, if the grayscale value of the pixel is greater than or equal to the grayscale threshold, marking the pixel as a subject pixel; if the grayscale value of the pixel is less than the grayscale threshold, marking the pixel as a background pixel; after marking, all background pixels in the grayscale thermal imaging image are removed to obtain a grayscale thermal imaging subject image.

[0009] Furthermore, performing image cropping and image feature processing on the first thermal imaging sample data to obtain the second thermal imaging sample data further includes the following sub-steps: Perform image feature processing on all grayscale thermal imaging subject images in the thermal imaging subject sample data, and obtain second thermal imaging sample data after completion; The image feature processing includes: for any pixel point in the grayscale thermal imaging main image, denoted as Ai, i represents the serial number of the pixel point, obtain the grayscale value of Ai and the grayscale values ​​of the pixels in the 8 neighborhoods around Ai, if there is a neighborhood without pixels in the 8 neighborhoods around Ai, then the grayscale value of the neighborhood without pixels is set to the grayscale value of Ai; calculate the grayscale difference between the grayscale value of Ai and the grayscale value of the pixels in the 8 neighborhoods around Ai respectively, and perform weighted average of the obtained 8 grayscale differences to obtain the average grayscale difference, which is marked as the feature enhancement value of Ai; based on the feature enhancement value of Ai, use the enhancement mapping formula to calculate the grayscale enhancement value of Ai, and the enhancement mapping formula is as follows: , where ZAi is the grayscale enhancement value of Ai, IAi is the grayscale value of Ai, c is the scaling factor, and EAi is the feature enhancement value of Ai; the grayscale enhancement value of Ai is used to replace the grayscale value of Ai, and the grayscale values ​​of all pixels in the grayscale thermal imaging subject image are repeatedly replaced to obtain the thermal imaging subject enhanced image.

[0010] Furthermore, constructing an original detection model and using the second thermal imaging sample data for training to obtain an alloy steel wire rope detection model includes the following sub-steps: The original detection model is constructed based on convolutional neural network and recurrent neural network. The original prediction model includes: CNN unit, intermediate connection layer and RNN unit. The CNN unit includes: CNN input layer, convolution processing layer and pooling layer. The RNN unit includes: RNN input layer, LSTM layer and output layer; the convolution processing layer includes G1 convolution layers.

[0011] Furthermore, constructing the original detection model and using the second thermal imaging sample data for training to obtain the alloy steel wire rope detection model also includes the following sub-steps: Image normalization processing: adjusting the image size of all thermal imaging subject enhanced images in the second thermal imaging sample data to L1*L2, and normalizing the grayscale enhancement values ​​of all pixels, scaling all grayscale enhancement values ​​to [0, 1], and obtaining the third thermal imaging sample data after completion; The third thermal imaging sample data is divided into a model training data set and a model test data set in a ratio of 8:2; the initial learning rate of the model training is set to f1, the training round is set to f2, the batch size is set to f3, and the learning rate is reduced by 5% after every f4 training rounds; the original detection model is trained using the model training data set.

[0012] Furthermore, constructing the original detection model and using the second thermal imaging sample data for training to obtain the alloy steel wire rope detection model also includes the following sub-steps: After the training of the original detection model is completed, the trained original detection model is marked as the first detection model; Based on the model test data set, the model test data set is input into the first detection model for model testing, and the recognition accuracy of the first detection model for the alloy steel wire rope with no defects, surface defects, broken wire defects and internal defects is calculated respectively; the recognition accuracy Y1 of the first detection model for no defects, the recognition accuracy Y2 of surface defects, the recognition accuracy Y3 of broken wire defects, and the recognition accuracy Y4 of internal defects are obtained; Set the appropriate threshold for no defects to Z1, the appropriate threshold for surface defects to Z2, the appropriate threshold for broken wire defects to Z3, and the appropriate threshold for internal defects to Z4; determine whether the first detection model satisfies Y1≥Z1 and Y2≥Z2 and Y3≥Z4 and Y4≥Z4; if satisfied, the first detection model is determined to be qualified, and the first detection model is marked as an alloy steel wire rope detection model; if not satisfied, the first detection model is determined to be unqualified, and the first detection model is re-trained using the model training data set. After the re-training is completed, it is determined again whether it is qualified until the alloy steel wire rope detection model is obtained.

[0013] Furthermore, based on the alloy steel wire rope detection model, the flaw detection of the alloy steel wire rope using the collected thermal imaging image set of the alloy steel wire rope includes the following sub-steps: Performing heating and cooling treatment on the alloy steel wire rope to be inspected, and collecting thermal imaging images of the alloy steel wire rope, and recording multiple grayscale thermal imaging images collected at the same rotation angle as a first image group, a second image group, and a third image group; The first image group, the second image group and the third image group are subjected to image cropping processing, image feature processing and image normalization processing in sequence, and the corresponding first test group, the second test group and the third test group are obtained after the processing is completed; The first test group, the second test group and the third test group are respectively input into the alloy steel wire rope detection model to obtain the flaw detection results of the alloy steel wire rope.

[0014] In a second aspect, the present application provides a flaw detection system for alloy steel wire rope production, including a data acquisition module, an image processing module, a model building module, and a flaw detection module; The data acquisition module includes a heating and cooling unit and an image acquisition unit, the heating and cooling unit is used to perform heating and cooling treatment on the alloy steel wire rope, and the image acquisition unit is used to acquire a thermal imaging image of the alloy steel wire rope to obtain first thermal imaging sample data; The image processing module includes a cropping processing unit and a feature processing unit, the cropping processing unit is used to perform image cropping processing on the first thermal imaging sample data, and the feature processing unit is used to perform image feature processing to obtain the second thermal imaging sample data; The model building module is used to build an original detection model and use the second thermal imaging sample data for training to obtain an alloy steel wire rope detection model; The flaw detection module is based on the alloy steel wire rope detection model and uses the collected thermal imaging image set of the alloy steel wire rope to perform flaw detection on the alloy steel wire rope.

[0015] Beneficial effects of the present invention: The present invention obtains first thermal imaging sample data by heating and cooling the alloy steel wire rope and collecting thermal imaging images of the alloy steel wire rope; performs image cropping and image feature processing on the first thermal imaging sample data to obtain second thermal imaging sample data; constructs an original detection model and uses the second thermal imaging sample data for training to obtain an alloy steel wire rope detection model; based on the alloy steel wire rope detection model, uses the collected thermal imaging image set of the alloy steel wire rope to perform flaw detection on the alloy steel wire rope; by capturing the dynamic feature information existing in the temperature change process of the steel wire rope, those defect features that will only appear obviously in the temperature change process can be effectively captured; The present invention heats the alloy steel wire rope and continuously collects thermal images during the cooling process. The advantage is that the complete thermal change process of the steel wire rope from the heated state to the gradual cooling can be obtained; this dynamic temperature change information contains rich details of the internal structure and surface state of the steel wire rope, more comprehensively reflects the characteristics of the steel wire rope, and improves the depth and accuracy of flaw detection; the grayscale values ​​of all pixels in the grayscale thermal imaging main image are replaced and enhanced through image feature processing. The advantage is that the grayscale difference between the defective part of the steel wire rope and the normal part can be improved to form a clear defect feature, thereby improving the detection efficiency. At the same time, due to the simplicity of the method, the complex calculation process is avoided and the processing complexity is reduced; the original detection model is constructed based on the convolutional neural network and the recurrent neural network, which can better understand and analyze the complex thermal phenomena in the temperature drop process and improve the detection capability and accuracy of defects. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 is a functional block diagram of the system of the present invention; Figure 2 is a flow chart of the steps of the method of the present invention; Figure 3 This is a schematic diagram of collecting grayscale thermal imaging images of the present invention; Figure 4 It is a schematic diagram of the structure of the original prediction model of the present invention; Figure 5 It is a schematic structural diagram of the electronic device of the present invention. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] Example 1, please refer to Figure 1 As shown, the present application provides a flaw detection system for alloy steel wire rope production, including a data acquisition module, an image processing module, a model building module and a flaw detection module; The data acquisition module includes a heating and cooling unit and an image acquisition unit. The heating and cooling unit is used to perform heating and cooling treatment on the alloy steel wire rope. The image acquisition unit is used to acquire a thermal imaging image of the alloy steel wire rope to obtain first thermal imaging sample data. The heating and cooling unit is equipped with a heating and cooling strategy, which includes: obtaining a certain number of alloy steel wire ropes, including defective alloy steel wire ropes and non-defective alloy steel wire ropes, and dividing the obtained alloy steel wire ropes into groups according to whether they have defects and the types of defects, including: non-defect group, surface defect group, broken wire defect group and internal defect group; surface defects include surface wear, corrosion and surface cracks; internal defects include internal cracks and internal voids; Heating and cooling process: the heating temperature is set to T0, and in this embodiment, the heating temperature T0 is 80 degrees Celsius; The alloy steel wire ropes of the defect-free group, the surface defect group, the broken wire defect group and the internal defect group are heated by a heating device respectively, and the heated alloy steel wire ropes are placed in a uniform standard environment for natural cooling; in this embodiment, the standard environment is a windless space of 10 degrees Celsius; when heating the alloy steel wire rope, the temperature should not be too high, generally 60-150 degrees Celsius; within this temperature range, the material properties of the steel wire rope are relatively stable, and no additional thermal stress or structural changes will be generated due to excessive temperature, which will affect the flaw detection results, and the heat flow inside the steel wire rope will be more obvious. For areas with internal cracks, holes and other defects, the hindering effect of heat conduction will cause more significant differences in the surface temperature distribution, making it easier to be detected by a thermal imager; The image acquisition unit is configured with an image acquisition strategy, which includes: in the natural cooling process, setting the time window to t1, collecting the grayscale thermal imaging image of the alloy steel wire rope at a first time interval within the time window, and recording the acquisition time and number of acquisitions for each acquisition, and classifying and storing according to the group corresponding to the alloy steel wire rope, marking it as the first thermal imaging sample data, and the first time interval is t2; in this embodiment, the time window t1 is 10 minutes, and the first time interval t2 is 10 seconds; grayscale thermal imaging is a presentation method of thermal imaging technology; it converts the infrared radiation emitted by an object into a black and white grayscale image to display the temperature distribution on the surface of the object; in the grayscale thermal imaging image, different grayscale values ​​represent different temperatures; Grayscale thermal imaging images of alloy steel wire ropes include: Figure 3As shown, keep the alloy steel wire rope unbent, place the alloy steel wire rope parallel to the lens plane of the thermal imaging device, collect the first grayscale thermal imaging image, and record the rotation angle of the alloy steel wire rope at this time as the initial angle of 0°, keep the alloy steel wire rope and the thermal imaging device still, rotate the alloy steel wire rope 120° clockwise with the rope direction as the central axis, collect the second grayscale thermal imaging image, and then rotate the alloy steel wire rope 120° clockwise with the rope direction as the central axis, collect the third grayscale thermal imaging image, and collect the grayscale thermal imaging images according to the number of collections and the collection order. Marking is performed, and multiple grayscale thermal imaging images acquired at the same rotation angle are merged and stored, that is, each alloy steel wire rope contains three rotation angles, and each rotation angle contains multiple grayscale thermal imaging images acquired at the first time interval within the time window; through rotation acquisition, it can ensure that the entire circumferential surface of the steel wire rope is inspected, and any area that may have defects can be avoided from being missed. For internal defects, such as internal cracks or inclusions, thermal imaging acquisition from different angles can provide richer information; three thermal imaging devices can also be placed around the steel wire rope for acquisition; In the specific implementation process, after the alloy steel wire rope is heated, the heat will be conducted inside and on its surface; under normal circumstances, the heat will be transferred in the steel wire rope in a relatively uniform manner, and will also show relatively regular changes during the temperature drop process; however, when there are defects, the heat conduction process will be disturbed, and the thermal changes in the defective part will be different from the thermal changes in the surrounding normal parts; for internal defects, such as internal cracks or voids, the conduction of heat at the defects will be blocked; surface defects, such as wear, corrosion and surface cracks, will change the thermal emissivity and thermal conductivity characteristics of the steel wire rope surface; broken wire defects will form new surface boundaries at the broken wires, and their thermal conduction and thermal emission conditions are different from those of intact steel wires; and by continuously collecting infrared images during the temperature drop process, the complete thermal change process of the steel wire rope from the high temperature state after heating to the gradual cooling can be recorded; this dynamic temperature change information can more comprehensively reflect the internal structural characteristics of the steel wire rope, and help to detect internal defects more accurately.

[0019] The image processing module includes a cropping processing unit and a feature processing unit, the cropping processing unit is used to perform image cropping processing on the first thermal imaging sample data, and the feature processing unit is used to perform image feature processing to obtain second thermal imaging sample data; The cropping processing unit is configured with an image cropping strategy, which includes: performing image cropping processing on all grayscale thermal imaging images in the first thermal imaging sample data, and obtaining thermal imaging subject sample data after the image cropping processing is completed; The image cropping process includes: setting the grayscale threshold to H0, which needs to be set according to the specific application scenario, and can generally be set to a grayscale value corresponding to a temperature 4-5 degrees Celsius higher than the background environment; traversing all the pixels in the grayscale thermal imaging image, if the grayscale value of the pixel is greater than or equal to the grayscale threshold, marking the pixel as the main pixel, that is, the pixel corresponding to the alloy steel wire rope; if the grayscale value of the pixel is less than the grayscale threshold, marking the pixel as the background pixel, that is, the pixel corresponding to the background environment; after marking, all the background pixels in the grayscale thermal imaging image are removed to obtain the main grayscale thermal imaging image; The feature processing unit is configured with a feature enhancement strategy, which includes: performing image feature processing on all grayscale thermal imaging subject images in the thermal imaging subject sample data, and obtaining second thermal imaging sample data after completion; The image feature processing includes: for any pixel point in the grayscale thermal imaging main image, record it as Ai, i represents the serial number of the pixel point, for example, A1 represents the first pixel, and in the order from left to right and from top to bottom, the pixel point in the upper left corner is the first pixel; obtain the grayscale value of Ai and the grayscale values ​​of the pixels in the 8 neighborhoods around Ai, that is, the grayscale values ​​of the pixels in the 8 directions of Ai directly above, upper right, directly right, lower right, directly below, directly below, directly below, directly left and upper left; If there is a neighborhood without pixels in the 8 neighborhoods around Ai, that is, the pixels at the edge of the image, for example, if A1 is located in the upper left corner, there are no pixels just above, right above, left below, left above, and left above A1; then the grayscale value of the neighborhood without pixels is set to be the grayscale value of Ai; this is to prevent the subsequent calculation of grayscale differences from being affected; the grayscale difference between the grayscale value of Ai and the grayscale values ​​of the 8 neighborhood pixels around Ai is calculated respectively, and the 8 grayscale differences are weighted averaged to obtain the average grayscale difference, which is marked as the feature enhancement value of Ai; For example, the grayscale value of A89 is 89; the grayscale values ​​of the adjacent pixels in the eight directions of A89, namely, 95, 94, 89, 88, 88, 89, 89, and 96 respectively; then the grayscale differences between A89 and the adjacent pixels in these eight directions are -15, -14, 0, 1, 1, 0, 0, and -16 respectively, setting the weights of the upper, right, lower, and left directions. The weight of the upper right, lower right, lower left and upper left is 0.1, then the average grayscale difference = 0.15*(-15+0+1+0)+0.1*[-14+1+0+(-16)]=-5; the setting of weights is based on the characteristic directions in which defects may appear in the image; for example, the crack defect of the wire rope appears linear in the image and is mostly horizontal and vertical, so the grayscale difference weights in the vertical and horizontal directions can be set higher; Based on the feature enhancement value of Ai, the grayscale enhancement value of Ai is calculated using the enhancement mapping formula. The enhancement mapping formula is as follows: , where ZAi is the grayscale enhancement value of Ai, IAi is the grayscale value of Ai, and c is the scaling factor. The scaling factor is mainly used to control the intensity of the feature enhancement value when it is superimposed on the original image; it plays a role in balancing the original image information and the enhanced feature information; since the size range of the calculated feature enhancement value may not match the pixel value range of the original image, the enhancement value can be adjusted to a suitable range through the scaling factor, so that it can effectively enhance the target features in the image, while avoiding excessive enhancement that causes image distortion or the original information to be covered; in this embodiment, c is 0.6; for example, the grayscale value of A89 is 89, and the feature enhancement value of A89 is -5, then ZA89=89+0.6*(-5)=86; EAi is the feature enhancement value of Ai; the grayscale value of Ai is replaced by the grayscale enhancement value of Ai, and the grayscale values ​​of all pixels in the grayscale thermal imaging subject image are repeatedly replaced to obtain the thermal imaging subject enhanced image; In the specific implementation process: the grayscale value of the thermal imaging image is related to the temperature, and the grayscale difference can reflect the temperature difference; by calculating the grayscale difference between each pixel and its adjacent pixels, the temperature change information in the image can be captured; for the thermal imaging image of the defective wire rope, there is a temperature difference between the defective area and the normal area, and this difference can be reflected in the grayscale difference calculation; then, by assigning weights to the grayscale differences in different directions, the contrast between the defective area and the normal area can be effectively enhanced, making the defects more obvious in the image.

[0020] The model building module is used to build an original detection model and use the second thermal imaging sample data for training to obtain an alloy steel wire rope detection model; The model building module configures the right model building strategy. The model building strategy includes: building the original detection model based on the convolutional neural network and the recurrent neural network, see Figure 4 As shown, the original prediction model includes: a CNN unit, an intermediate connection layer and an RNN unit. The CNN unit includes: a CNN input layer, a convolution processing layer and a pooling layer. The RNN unit includes: an RNN input layer, an LSTM layer and an output layer. The convolution processing layer includes G1 convolution layers. In this embodiment, the convolution processing layer includes 2 convolution layers. Image normalization processing: adjusting the image size of all thermal imaging subject enhanced images in the second thermal imaging sample data to L1*L2. In this embodiment, L1*L2 is 224*224; performing normalization processing on the grayscale enhancement values ​​of all pixels, scaling all grayscale enhancement values ​​to [0, 1]. The normalization processing can obtain the maximum value of the grayscale enhancement value in the second thermal imaging sample data, and then dividing the grayscale enhancement values ​​of all other pixels by the maximum value, thereby scaling all grayscale enhancement values ​​to [0, 1]. After completion, the third thermal imaging sample data is obtained; The third thermal imaging sample data is divided into a model training data set and a model test data set in a ratio of 8:2; the initial learning rate of the model training is set to f1, the training round is set to f2, the batch size is set to f3, and the learning rate is reduced by 5% after every f4 training rounds; the original detection model is trained using the model training data set; in this embodiment, f1=0.005, f2=200, f3=32, f=10, that is, the learning rate is reduced by 5% after every 10 training rounds; after completing the training of the original detection model, the trained original detection model is marked as the first detection model; Based on the model test data set, the model test data set is input into the first detection model for model testing, and the recognition accuracy of the first detection model for the alloy steel wire rope with no defects, surface defects, broken wire defects and internal defects is calculated respectively, that is, the proportion of the number of normal recognitions in the corresponding defect types to the total number of recognitions; the recognition accuracy Y1 of the first detection model for no defects, the recognition accuracy Y2 of surface defects, the recognition accuracy Y3 of broken wire defects, and the recognition accuracy Y4 of internal defects are obtained; Set the appropriate threshold for no defects to Z1, the appropriate threshold for surface defects to Z2, the appropriate threshold for broken wire defects to Z3, and the appropriate threshold for internal defects to Z4; in this embodiment, Z1=Z2=Z3=0.9, Z4=0.85, because internal defects are more difficult to identify than other defects, the standard can be appropriately lowered; determine whether the first detection model satisfies Y1≥Z1 and Y2≥Z2 and Y3≥Z4 and Y4≥Z4; if satisfied, the first detection model is determined to be qualified, and the first detection model is marked as an alloy steel wire rope detection model; if not satisfied, the first detection model is determined to be unqualified, and the first detection model is retrained using the model training data set. The initial learning rate, training rounds, batch size and other model parameters can be appropriately changed in the secondary training to quickly obtain a qualified model; after the retraining is completed, determine whether it is qualified again until the alloy steel wire rope detection model is obtained; In the specific implementation process, the convolution layer of the convolutional neural network slides on the image through the convolution kernel to automatically extract the spatial features of the image; the pooling layer downsamples the feature map output by the convolution layer to reduce the amount of data while retaining the main spatial features; the recurrent neural network is used to process the temporal correlation of the image sequence; the recurrent neural network can receive the spatial features extracted by the convolutional neural network at each time step, and learn the dynamic feature changes during the temperature drop process based on the hidden state of the previous time step and the input of the current time step; for example, it can learn how the temperature gradually decreases over time and the temperature change law of the defect area during this process; by learning the temporal features of the entire image sequence, the recurrent neural network can better understand the dynamic behavior of the defect during the temperature drop process, and thus combine it with the spatial features extracted by the convolutional neural network to more accurately determine the type of defect; therefore, the time step of the RNN unit needs to be set according to the number of acquisitions in the time window, for example, if the number of acquisitions in the time window is 60, that is, 60 thermal images are collected at one rotation angle, then the time step should also be set to 60, so that the model can fully process the temporal information of the entire image sequence.

[0021] The flaw detection module is based on the alloy steel wire rope detection model and uses the collected thermal imaging image set of the alloy steel wire rope to perform flaw detection on the alloy steel wire rope; The flaw detection module is equipped with a flaw detection strategy, which includes: heating and cooling the alloy steel wire rope to be detected, collecting thermal imaging images of the alloy steel wire rope, and recording multiple grayscale thermal imaging images collected at the same rotation angle as the first image group, the second image group and the third image group, that is, three groups of images collected at rotation angles of 0°, 120° and 240°; The first image group, the second image group and the third image group are subjected to image cropping, image feature processing and image normalization processing in sequence, respectively, so that the pixel values ​​of the images can be unified into a standard range, so that the model can process the images on the same numerical scale, avoiding detection errors caused by scale differences; after completion, the corresponding first test group, the second test group and the third test group are obtained; Inputting the first test group, the second test group and the third test group into the alloy steel wire rope detection model respectively to obtain the flaw detection results of the alloy steel wire rope; In the specific implementation process, thermal imaging detection at a single angle may be interfered by other factors, so it is necessary to obtain thermal imaging images from multiple angles and input them into the model so that they can be compared and verified with each other; if the same defect is detected at multiple angles, such as an internal defect, then the possibility that the defect is a real defect is greatly increased, thereby effectively improving the accuracy and reliability of detection and reducing misjudgment.

[0022] Example 2, please refer to Figure 2 As shown, the present application provides a flaw detection method for alloy steel wire rope production, comprising the following steps: Step S1, heating and cooling the alloy steel wire rope, and collecting a thermal imaging image of the alloy steel wire rope to obtain first thermal imaging sample data; Step S1 includes the following sub-steps: Step S101, obtaining a certain number of alloy steel wire ropes, including defective alloy steel wire ropes and non-defective alloy steel wire ropes, and dividing the obtained alloy steel wire ropes into groups according to whether they have defects and defect types, and the groups include: a non-defect group, a surface defect group, a broken wire defect group, and an internal defect group; Step S102, heating and cooling treatment: setting the heating temperature to T0, using a heating device to heat the alloy steel wire ropes of the non-defect group, the surface defect group, the broken wire defect group and the internal defect group respectively, and placing the heated alloy steel wire ropes in a uniform standard environment for natural cooling; Step S103, during the natural cooling process, the time window is set to t1, and the grayscale thermal imaging image of the alloy steel wire rope is collected at a first time interval within the time window, and the collection time and the number of collections of each collection are recorded, and the images are classified and stored according to the groups corresponding to the alloy steel wire ropes, and marked as the first thermal imaging sample data, and the first time interval is t2; Step S104, collecting a grayscale thermal imaging image of the alloy steel wire rope; Step S1 includes the following sub-steps: Step S1041, keeping the alloy steel wire rope unbent, placing the alloy steel wire rope parallel to the lens plane of the thermal imaging device, capturing the first grayscale thermal imaging image, and recording the rotation angle of the alloy steel wire rope at this time as an initial angle of 0°; Step S1042, keeping the alloy steel wire rope and the thermal imaging device stationary, rotating the alloy steel wire rope 120° clockwise with the rope direction as the central axis, and collecting a second grayscale thermal imaging image; Step S1043, the alloy steel wire rope is rotated 120° clockwise with the rope direction as the central axis, and a third grayscale thermal imaging image is collected; Step S1044: Mark the collected grayscale thermal imaging images according to the number of collection times and the collection order, and merge and store multiple grayscale thermal imaging images collected at the same rotation angle.

[0023] Step S2, performing image cropping and image feature processing on the first thermal imaging sample data to obtain second thermal imaging sample data; Step S2 includes the following sub-steps: Step S201, performing image cropping processing on all grayscale thermal imaging images in the first thermal imaging sample data, and obtaining thermal imaging subject sample data after the cropping processing is completed; Step S202: image cropping processing; Step S202 includes the following sub-steps: Step S2021, setting the grayscale threshold to H0, traversing all pixels in the grayscale thermal imaging image, and if the grayscale value of a pixel is greater than or equal to the grayscale threshold, marking the pixel as a subject pixel; Step S2022: if the grayscale value of the pixel is less than the grayscale threshold, mark the pixel as a background pixel; after marking, remove all background pixels in the grayscale thermal imaging image to obtain a grayscale thermal imaging subject image; Step S203, performing image feature processing on all grayscale thermal imaging subject images in the thermal imaging subject sample data, and obtaining second thermal imaging sample data after completion; Step S204, image feature processing includes: Step S204 includes the following sub-steps: Step S2041, for any pixel point in the grayscale thermal imaging subject image, denoted as Ai, where i represents the sequence number of the pixel point, obtain the grayscale value of Ai and the grayscale values ​​of the pixels in the eight neighborhoods around Ai. If there is a neighborhood without a pixel point among the eight neighborhoods around Ai, the grayscale value of the neighborhood without a pixel point is set to be the grayscale value of Ai; Step S2042, respectively calculating the grayscale difference between the grayscale value of Ai and the grayscale values ​​of the eight neighboring pixels around Ai, and performing weighted average on the eight grayscale differences to obtain an average grayscale difference, which is marked as the feature enhancement value of Ai; Step S2043, based on the feature enhancement value of Ai, the grayscale enhancement value of Ai is calculated using the enhancement mapping formula, and the enhancement mapping formula is as follows: , where ZAi is the grayscale enhancement value of Ai, IAi is the grayscale value of Ai, c is the scaling factor, and EAi is the feature enhancement value of Ai; Step S2044, using the grayscale enhancement value of Ai to replace the grayscale value of Ai, repeatedly replacing the grayscale values ​​of all pixels in the grayscale thermal imaging subject image to obtain a thermal imaging subject enhanced image.

[0024] Step S3, constructing an original detection model, and using the second thermal imaging sample data for training to obtain an alloy steel wire rope detection model; Step S3 includes the following sub-steps: Step S301, constructing an original detection model based on a convolutional neural network and a recurrent neural network, wherein the original prediction model includes: a CNN unit, an intermediate connection layer, and an RNN unit, wherein the CNN unit includes: a CNN input layer, a convolution processing layer, and a pooling layer, and the RNN unit includes: an RNN input layer, an LSTM layer, and an output layer; and the convolution processing layer includes G1 convolution layers; Step S302, image normalization processing: adjusting the image size of all thermal imaging subject enhanced images in the second thermal imaging sample data to L1*L2, and normalizing the grayscale enhancement values ​​of all pixels, scaling all grayscale enhancement values ​​to [0, 1], and obtaining the third thermal imaging sample data after completion; Step S303, dividing the third thermal imaging sample data into a model training data set and a model testing data set in a ratio of 8:2; Step S304, setting the initial learning rate of the model training to f1, the training rounds to f2, the batch size to f3, and reducing the learning rate by 5% after each f4 training rounds; training the original detection model using the model training data set; Step S305, after the training of the original detection model is completed, the trained original detection model is marked as the first detection model; Step S306, based on the model test data set, input the model test data set into the first detection model for model testing, and respectively calculate the recognition accuracy of the first detection model for the alloy steel wire rope with no defects, surface defects, broken wire defects, and internal defects; obtain the recognition accuracy Y1 of the first detection model for no defects, the recognition accuracy Y2 of surface defects, the recognition accuracy Y3 of broken wire defects, and the recognition accuracy Y4 of internal defects; Step S307, setting the suitable threshold for no defect to Z1, the suitable threshold for surface defect to Z2, the suitable threshold for broken wire defect to Z3, and the suitable threshold for internal defect to Z4; judging whether the first detection model satisfies Y1≥Z1 and Y2≥Z2 and Y3≥Z4 and Y4≥Z4; Step S308: If the conditions are met, the first detection model is determined to be qualified and the first detection model is marked as an alloy steel wire rope detection model; Step S309: if not satisfied, the first detection model is determined to be unqualified, and the first detection model is retrained using the model training data set. After the retraining is completed, it is determined again whether it is qualified until the alloy steel wire rope detection model is obtained.

[0025] Step S4, based on the alloy steel wire rope detection model, using the collected thermal imaging image set of the alloy steel wire rope to perform flaw detection on the alloy steel wire rope; Step S4 includes the following sub-steps: Step S401, heating and cooling the alloy steel wire rope to be inspected, and collecting thermal imaging images of the alloy steel wire rope, and recording multiple grayscale thermal imaging images collected at the same rotation angle as a first image group, a second image group, and a third image group; Step S402, performing image cropping, image feature processing and image normalization processing on the first image group, the second image group and the third image group in sequence, and obtaining the corresponding first test group, the second test group and the third test group after completion; Step S403, inputting the first to-be-tested group, the second to-be-tested group and the third to-be-tested group into the alloy steel wire rope detection model respectively to obtain the flaw detection results of the alloy steel wire rope.

[0026] Example 3, please refer to Figure 5 As shown, Figure 5 The structural schematic diagram of an electronic device is illustrated, and the electronic device may include: a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other through the communication bus. The memory stores computer-readable instructions, and the processor can call the instructions in the memory. When the computer-readable instructions are executed by the processor, the steps in a flaw detection method for alloy steel wire rope production are executed to achieve the following functions: heating and cooling the alloy steel wire rope, and collecting thermal imaging images of the alloy steel wire rope to obtain first thermal imaging sample data; performing image cropping and image feature processing on the first thermal imaging sample data to obtain second thermal imaging sample data; constructing an original detection model, and using the second thermal imaging sample data for training to obtain an alloy steel wire rope detection model; based on the alloy steel wire rope detection model, using the collected thermal imaging image set of the alloy steel wire rope to perform flaw detection on the alloy steel wire rope.

[0027] In addition, the logic instructions in the above-mentioned memory can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0028] Embodiment 4, the present application also provides a computer-readable storage medium, the present application provides a storage medium, on which a computer program is stored. When the computer program is executed by the processor, the steps in the above-mentioned flaw detection method for alloy steel wire rope production are executed to achieve the following functions: heating and cooling the alloy steel wire rope, and collecting thermal imaging images of the alloy steel wire rope to obtain first thermal imaging sample data; performing image cropping and image feature processing on the first thermal imaging sample data to obtain second thermal imaging sample data; constructing an original detection model, and using the second thermal imaging sample data for training to obtain an alloy steel wire rope detection model; based on the alloy steel wire rope detection model, using the collected thermal imaging image set of the alloy steel wire rope to perform flaw detection on the alloy steel wire rope.

[0029] Through the description of the above implementation methods, the embodiments of the present invention can be provided as methods, systems or computer program products. Based on such an understanding, the above technical solutions can be essentially or partly contributed to the prior art in the form of software products, which can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and include several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0030] In the embodiments provided in the present application, it should be understood that the disclosed system or method can be implemented in other ways. The embodiments described above are merely schematic. For example, the division of modules or units is only a logical function division. There may be other division methods in actual implementation. For example, multiple modules or units can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of systems, modules and units can be electrical, mechanical or other forms.

[0031] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A flaw detection method for alloy steel wire rope production, characterized in that: The steps include: Performing heating and cooling treatment on the alloy steel wire rope, and collecting a thermal imaging image of the alloy steel wire rope to obtain first thermal imaging sample data; Performing image cropping and image feature processing on the first thermal imaging sample data to obtain second thermal imaging sample data; An original detection model is constructed and trained using the second thermal imaging sample data to obtain an alloy steel wire rope detection model; Based on the alloy steel wire rope detection model, the alloy steel wire rope is subjected to flaw detection using the collected thermal imaging image set of the alloy steel wire rope.

2. A flaw detection method for alloy steel wire rope production according to claim 1, characterized in that: Performing heating and cooling treatment on the alloy steel wire rope and collecting a thermal imaging image of the alloy steel wire rope to obtain first thermal imaging sample data includes the following sub-steps: A certain number of alloy steel wire ropes are obtained, including defective alloy steel wire ropes and non-defective alloy steel wire ropes, and the obtained alloy steel wire ropes are divided into groups according to whether there are defects and the types of defects, including: a non-defect group, a surface defect group, a broken wire defect group, and an internal defect group; Heating and cooling treatment: Set the heating temperature to T0, and use the heating device to heat the alloy steel wire ropes of the defect-free group, surface defect group, broken wire defect group and internal defect group respectively, and place the heated alloy steel wire ropes in a unified standard environment for natural cooling.

3. A flaw detection method for alloy steel wire rope production according to claim 2, characterized in that: The step of heating and cooling the alloy steel wire rope and collecting a thermal imaging image of the alloy steel wire rope to obtain first thermal imaging sample data further includes the following sub-steps: During the natural cooling process, the time window is set to t1, and the grayscale thermal imaging image of the alloy steel wire rope is collected at the first time interval within the time window, and the collection time and the number of collections of each collection are recorded, and the images are classified and stored according to the groups corresponding to the alloy steel wire ropes, marked as the first thermal imaging sample data, and the first time interval is t2; Collecting the grayscale thermal imaging image of the alloy steel wire rope includes: keeping the alloy steel wire rope unbent, placing the alloy steel wire rope parallel to the lens plane of the thermal imaging device, collecting the first grayscale thermal imaging image, and recording the rotation angle of the alloy steel wire rope at this time as the initial angle of 0°, keeping the alloy steel wire rope and the thermal imaging device stationary, rotating the alloy steel wire rope 120° clockwise with the rope direction as the central axis, collecting the second grayscale thermal imaging image, and then rotating the alloy steel wire rope 120° clockwise with the rope direction as the central axis, collecting the third grayscale thermal imaging image, marking the collected grayscale thermal imaging images according to the number of collections and the collection order, and merging and storing multiple grayscale thermal imaging images collected at the same rotation angle.

4. A flaw detection method for alloy steel wire rope production according to claim 3, characterized in that: Performing image cropping and image feature processing on the first thermal imaging sample data to obtain second thermal imaging sample data includes the following sub-steps: Perform image cropping processing on all grayscale thermal imaging images in the first thermal imaging sample data, and obtain thermal imaging subject sample data after the processing is completed; The image cropping process includes: setting the grayscale threshold to H0, traversing all pixels in the grayscale thermal imaging image, if the grayscale value of the pixel is greater than or equal to the grayscale threshold, marking the pixel as a subject pixel; if the grayscale value of the pixel is less than the grayscale threshold, marking the pixel as a background pixel; after marking, all background pixels in the grayscale thermal imaging image are removed to obtain a grayscale thermal imaging subject image.

5. A flaw detection method for alloy steel wire rope production according to claim 4, characterized in that: Performing image cropping and image feature processing on the first thermal imaging sample data to obtain the second thermal imaging sample data also includes the following sub-steps: Performing image feature processing on all grayscale thermal imaging subject images in the thermal imaging subject sample data, and obtaining second thermal imaging sample data after completion; The image feature processing includes: for any pixel point in the grayscale thermal imaging main image, denoted as Ai, i represents the serial number of the pixel point, obtain the grayscale value of Ai and the grayscale values ​​of the pixels in the 8 neighborhoods around Ai, if there is a neighborhood without pixels in the 8 neighborhoods around Ai, then the grayscale value of the neighborhood without pixels is set to the grayscale value of Ai; calculate the grayscale difference between the grayscale value of Ai and the grayscale value of the pixels in the 8 neighborhoods around Ai respectively, and perform weighted average of the obtained 8 grayscale differences to obtain the average grayscale difference, which is marked as the feature enhancement value of Ai; based on the feature enhancement value of Ai, use the enhancement mapping formula to calculate the grayscale enhancement value of Ai, and the enhancement mapping formula is as follows: , where ZAi is the grayscale enhancement value of Ai, IAi is the grayscale value of Ai, c is the scaling factor, and EAi is the feature enhancement value of Ai; the grayscale enhancement value of Ai is used to replace the grayscale value of Ai, and the grayscale values ​​of all pixels in the grayscale thermal imaging subject image are repeatedly replaced to obtain the thermal imaging subject enhanced image.

6. A flaw detection method for alloy steel wire rope production according to claim 5, characterized in that: Constructing the original detection model and using the second thermal imaging sample data for training to obtain the alloy steel wire rope detection model includes the following sub-steps: The original detection model is constructed based on convolutional neural network and recurrent neural network. The original prediction model includes: CNN unit, intermediate connection layer and RNN unit. The CNN unit includes: CNN input layer, convolution processing layer and pooling layer. The RNN unit includes: RNN input layer, LSTM layer and output layer; the convolution processing layer includes G1 convolution layers.

7. A flaw detection method for alloy steel wire rope production according to claim 6, characterized in that: Constructing the original detection model and using the second thermal imaging sample data for training to obtain the alloy steel wire rope detection model also includes the following sub-steps: Image normalization processing: adjusting the image size of all thermal imaging subject enhanced images in the second thermal imaging sample data to L1*L2, and normalizing the grayscale enhancement values ​​of all pixels, scaling all grayscale enhancement values ​​to [0, 1], and obtaining the third thermal imaging sample data after completion; The third thermal imaging sample data is divided into a model training data set and a model test data set in a ratio of 8:2; the initial learning rate of the model training is set to f1, the training round is set to f2, the batch size is set to f3, and the learning rate is reduced by 5% after each f4 training rounds; The original detection model is trained using the model training dataset.

8. A flaw detection method for alloy steel wire rope production according to claim 7, characterized in that: Constructing the original detection model and using the second thermal imaging sample data for training to obtain the alloy steel wire rope detection model also includes the following sub-steps: After the training of the original detection model is completed, the trained original detection model is marked as the first detection model; Based on the model test data set, the model test data set is input into the first detection model for model testing, and the recognition accuracy of the first detection model for the alloy steel wire rope with no defects, surface defects, broken wire defects and internal defects is calculated respectively; Obtain the first detection model's recognition accuracy rate Y1 for no defects, the recognition accuracy rate Y2 for surface defects, the recognition accuracy rate Y3 for broken wire defects, and the recognition accuracy rate Y4 for internal defects; Set the appropriate threshold for no defects to Z1, the appropriate threshold for surface defects to Z2, the appropriate threshold for broken wire defects to Z3, and the appropriate threshold for internal defects to Z4; determine whether the first detection model satisfies Y1≥Z1 and Y2≥Z2 and Y3≥Z4 and Y4≥Z4; if satisfied, the first detection model is determined to be qualified, and the first detection model is marked as an alloy steel wire rope detection model; if not satisfied, the first detection model is determined to be unqualified, and the first detection model is re-trained using the model training data set. After the re-training is completed, it is determined again whether it is qualified until the alloy steel wire rope detection model is obtained.

9. A flaw detection method for alloy steel wire rope production according to claim 8, characterized in that: Based on the alloy steel wire rope detection model, the flaw detection of the alloy steel wire rope using the collected thermal imaging image set of the alloy steel wire rope includes the following sub-steps: Performing heating and cooling treatment on the alloy steel wire rope to be inspected, and collecting thermal imaging images of the alloy steel wire rope, and recording multiple grayscale thermal imaging images collected at the same rotation angle as a first image group, a second image group, and a third image group; The first image group, the second image group and the third image group are subjected to image cropping processing, image feature processing and image normalization processing in sequence, and the corresponding first test group, the second test group and the third test group are obtained after the processing is completed; The first test group, the second test group and the third test group are respectively input into the alloy steel wire rope detection model to obtain the flaw detection results of the alloy steel wire rope.

10. A flaw detection system for alloy steel wire rope production, used to implement a flaw detection method for alloy steel wire rope production as claimed in any one of claims 1 to 9, characterized in that: It includes data acquisition module, image processing module, model building module and flaw detection module; The data acquisition module includes a heating and cooling unit and an image acquisition unit, the heating and cooling unit is used to perform heating and cooling treatment on the alloy steel wire rope, and the image acquisition unit is used to acquire a thermal imaging image of the alloy steel wire rope to obtain first thermal imaging sample data; The image processing module includes a cropping processing unit and a feature processing unit, the cropping processing unit is used to perform image cropping processing on the first thermal imaging sample data, and the feature processing unit is used to perform image feature processing to obtain the second thermal imaging sample data; The model building module is used to build an original detection model and use the second thermal imaging sample data for training to obtain an alloy steel wire rope detection model; The flaw detection module is based on the alloy steel wire rope detection model and uses the collected thermal imaging image set of the alloy steel wire rope to perform flaw detection on the alloy steel wire rope.

Images