Extrusion forming detection method of copper piece extruder

Abstract

The invention relates to the technical field of image processing, in particular to a copper part extruder extrusion forming detection method which comprises the steps that multiple video frames in the copper part extrusion process are collected, and a stationary moment set of any position point of a copper part is calculated; initializing an initial section of any position point, responding to the situation that the intersection of the initial section stationary moment set and the initial section adjacent position point stationary moment set is not empty, adding the adjacent position points into the initial section until the intersection of the stationary moment sets is empty, and obtaining a stationary section of the position points; according to the stationary moment set of the stationary segment, splicing neighborhood images of all position points in the stationary segment to obtain an extrusion image segment of the stationary segment; and splicing the extruded image segments to obtain a detection result. According to the technical scheme, the influence of vibration on extrusion forming detection can be eliminated, and an accurate detection result is obtained.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to a method for detecting extrusion molding of a copper extruder. Background Art

[0002] The copper extruder is a mechanical device used to produce copper parts such as copper wire, copper wire or copper rod by extruding copper material from the die orifice under high pressure to form profiles or parts with specific cross-sectional shapes. In the process of using the copper extruder to extrude copper material, copper parts will have defects such as uneven thickness, surface cracks or even breakage. In order to ensure the quality of copper extrusion molding, the extrusion process of the copper extruder needs to be tested.

[0003] At present, a patent application document with publication number CN118446946A discloses a method and system for intelligent control of titanium wire drawing process, wherein the method includes: obtaining a grayscale image of the titanium wire to obtain edge pixels; according to a preset similarity measurement standard, taking the number of similar pixels of each edge pixel as the total target number corresponding to each edge pixel; taking other edge pixels on a straight line in the gradient direction of the target point as gradient pixels, and the straight line passes through the target point, and determining the edge degree of the titanium wire according to the total target number of the target point and the difference between the total target number of the target point and the gradient pixel point, thereby obtaining a target gamma value, and obtaining an updated grayscale value of the target point based on a gamma transform, thereby obtaining an enhanced image, thereby detecting defects, and adjusting process parameters.

[0004] The above method enhances the brightness of the grayscale image and detects metal wire defects in the wire drawing process based on the enhanced image. However, during the extrusion molding process, the vibration of the copper extruder will cause copper wires, copper wires and other copper parts to vibrate, which in turn causes a large amount of noise in the grayscale image. Brightness enhancement cannot eliminate the noise caused by the vibration, and the extrusion molding detection result is inaccurate. Summary of the invention

[0005] In order to solve the technical problem of inaccurate extrusion molding detection results, the present application provides a copper extrusion machine extrusion molding detection method, which can eliminate the influence of vibration on extrusion molding detection and obtain accurate detection results.

[0006] In a first aspect of the present application, a method for detecting extrusion molding of a copper part extruder is provided, the method comprising: collecting multiple video frames of a copper part extrusion process, and intercepting a neighborhood image of an arbitrary position point in each video frame according to an extrusion speed; calculating a difference map between an arbitrary neighborhood image and two adjacent neighborhood images, taking the maximum value of the average pixel value in the difference map as the vibration amplitude of the position point at the corresponding timestamp of the neighborhood image, and taking the timestamp with a vibration amplitude less than a preset amplitude as a stable moment, to obtain a stable moment set of each position point; initializing an initial segment of an arbitrary position point, in response to the intersection of the initial segment stable moment set and the stable moment set of adjacent position points of the initial segment being not empty, adding the adjacent position points to the initial segment until the intersection of the stable moment sets is empty, to obtain a stable segment of the position point, the initial segment stable moment set being the intersection of the stable moment sets of each position point in the initial segment; splicing the neighborhood images of each position point in the stable segment according to the stable moment set of the stable segment, to obtain an extruded image segment of the stable segment; splicing each extruded image segment to obtain an extruded image, and obtaining a detection result according to the extruded image.

[0007] The video frames in the copper part extrusion process are collected. For any position point in the copper part, the pixel coordinates corresponding to the position point in each video frame are located according to the extrusion speed of the copper part and the acquisition frame rate of the video frame, and the video frame is intercepted according to the pixel coordinates to obtain the neighborhood image of the position point in each video frame; further, the vibration amplitude of each time stamp is calculated according to the difference map between the neighborhood image and two adjacent neighborhood images to obtain the stable moment set of each position point; the position point is used as the initial segment, and the initial segment is continuously expanded according to the intersection of the position point and the stable moment set of the adjacent position points until the stable segment of each position point in the copper part is obtained, and multiple position points in a stable segment can be in a stable state at the same time stamp; the neighborhood images of each position point in the stable segment are spliced ​​according to the stable moment set of the stable segment to obtain the extrusion image segment of the stable segment, which is the image information of the copper part in the stable state, and can eliminate the noise generated by vibration in the video frame, thereby improving the accuracy of the extrusion molding detection result; each extrusion image segment is spliced ​​to obtain the extrusion image, and accurate detection results are obtained based on the extrusion image.

[0008] Preferably, intercepting a neighborhood image of an arbitrary position point in each video frame according to the extrusion speed includes: determining the pixel coordinates of the position point in any video frame according to the extrusion speed, intercepting the video frame according to a preset number of pixels on both sides of the pixel coordinates, and obtaining a neighborhood image of the position point in the video frame.

[0009] Preferably, the location point Timestamp The vibration amplitude for: , , and Location points In timestamp , Timestamp and timestamp The neighborhood image of Indicates calculating the average pixel value.

[0010] The larger the average pixel value in the difference image, the more relative motion occurs between the two adjacent neighborhood images. The maximum value of the average pixel value in the two difference images corresponding to a neighborhood image is taken as the vibration amplitude of the location point, and the vibration amplitude of the location point at the corresponding timestamp of the neighborhood image is accurately quantified.

[0011] Preferably, the initialization of the initial segment of the arbitrary position point includes: the initial segment includes the arbitrary position point.

[0012] Preferably, obtaining the extruded image segment of the stable segment includes: obtaining a video frame with an arbitrary time stamp in the stable moment set of the stable segment, intercepting the video frame according to the pixel point coordinates of each position point in the stable segment in the video frame, and obtaining the initial image segment of the stable segment; and taking the average of all the initial image segments as the extruded image segment of the stable segment.

[0013] A stable moment set of a stable segment includes at least one timestamp, and one timestamp corresponds to an initial image segment. The initial image segment is image information of a stable segment in a copper part in a stable state, and can eliminate noise caused by vibration in a video frame.

[0014] Preferably, the method of taking the mean of all initial image segments as the extruded image segment of the stable segment includes: obtaining the pixel value of an arbitrary position point in each initial image segment to obtain a pixel value sequence; determining the weighting coefficient of each initial image segment according to the vibration amplitude of the position point at the time stamp corresponding to the initial image segment, wherein the weighting coefficient is negatively correlated with the vibration amplitude; performing weighted averaging on the pixel value sequence according to the weighting coefficient to obtain the weighted pixel value of the position point, wherein the weighted pixel values ​​of all position points in the stable segment constitute the extruded image segment.

[0015] Since the vibration amplitudes of various positions in the stable segment at different timestamps are different, the mean of all initial image segments is directly used as the extruded image segment, ignoring the difference in vibration amplitudes of various positions in the stable segment at different timestamps. As a result, the extruded image segment is greatly affected by the vibration amplitude, and the image information of the copper part in the stable state cannot be accurately obtained. Therefore, in the process of calculating the mean of all initial image segments, the vibration amplitudes of various positions in different initial image segments, that is, at different timestamps, can be considered to further eliminate the noise caused by vibration in the video frame.

[0016] Preferably, the location point In timestamp The weighting factor for: , For location point In timestamp The vibration amplitude, For location point The sum of the vibration amplitudes at corresponding timestamps in all initial image segments.

[0017] Preferably, the stitching of the extruded image segments comprises: calculating a difference map of overlapping areas of any two extruded image segments, updating the translation parameters of each extruded image segment, and completing the stitching of the two extruded image segments when the average pixel value of the difference map reaches a minimum value.

[0018] Preferably, obtaining the detection result based on the extrusion image includes: dividing the extrusion image into multiple image blocks along the extrusion direction of the copper part, calculating the similarity between any image blocks to construct a similarity matrix; calculating the Euclidean distance between the similarity matrix and the unit matrix, and in response to the Euclidean distance being greater than a preset threshold, the detection result is abnormal quality, otherwise, the detection result is normal quality.

[0019] In response to the Euclidean distance being greater than the preset threshold, it indicates that there are large differences in the image features of the copper parts between the image blocks, and defects such as uneven thickness, surface cracks, and even fractures appear on the surface of the copper parts. At this time, the detection result is abnormal quality, thereby realizing quality detection during the extrusion process of the copper parts.

[0020] Preferably, in response to failure to obtain the extrusion image, an early warning message indicating that the copper extruder has an abnormal operating state is issued.

[0021] If any position point is in a state of large vibration amplitude during the copper extrusion process, there will be no overlapping area between the two extrusion image segments, and splicing will be impossible, that is, the complete extrusion image cannot be obtained, indicating that the copper extruder is in a state of continuous and violent vibration, and the copper extruder is in an abnormal working state.

[0022] The technical solution of this application has the following beneficial technical effects: The video frames in the copper part extrusion process are collected. For any position point in the copper part, the pixel coordinates corresponding to the position point in each video frame are located according to the extrusion speed of the copper part and the acquisition frame rate of the video frame, and the video frame is intercepted according to the pixel coordinates to obtain the neighborhood image of the position point in each video frame; further, the vibration amplitude of each time stamp is calculated according to the difference map between the neighborhood image and two adjacent neighborhood images to obtain the stable moment set of each position point; the position point is used as the initial segment, and the initial segment is continuously expanded according to the intersection of the position point and the stable moment set of the adjacent position points until the stable segment of each position point in the copper part is obtained, and multiple position points in a stable segment can be in a stable state at the same time stamp; the neighborhood images of each position point in the stable segment are spliced ​​according to the stable moment set of the stable segment to obtain the extrusion image segment of the stable segment, which is the image information of the copper part in the stable state, and can eliminate the noise generated by vibration in the video frame, thereby improving the accuracy of the extrusion molding detection result; each extrusion image segment is spliced ​​to obtain the extrusion image, and accurate detection results are obtained based on the extrusion image. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is a flow chart of a copper part extrusion machine extrusion molding detection method according to an embodiment of the present application. DETAILED DESCRIPTION

[0024] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments of the present application, all other embodiments obtained by technicians in the field without creative work are within the scope of protection of the present application.

[0025] According to a first aspect of the present application, the present application provides a method for detecting extrusion molding of a copper part extrusion machine. Figure 1 1 is a flow chart of a copper extrusion machine extrusion molding detection method according to an embodiment of the present application. Figure 1 As shown, the copper extruder extrusion molding detection method includes steps S101 to S104, which are described in detail below.

[0026] S101, collecting multiple video frames of the copper part extrusion process, intercepting the neighborhood image of any position point in each video frame according to the extrusion speed; calculating the difference map between any neighborhood image and two adjacent neighborhood images, taking the maximum value of the average pixel value in the difference map as the vibration amplitude of the position point at the corresponding timestamp of the neighborhood image, taking the timestamp with a vibration amplitude less than a preset amplitude as the stable moment, and obtaining the stable moment set of each position point.

[0027] In one embodiment, a copper extruder is used to extrude copper material placed at one end of a die orifice to the other end of the die orifice, thereby extruding the copper material into copper products such as copper wire or copper rod with a preset cross-sectional area. At the same time, at the other end of the die orifice, a traction device will pull the copper wire or copper rod to move at a preset extrusion speed to ensure that the copper wire or copper rod will continue to be extruded at the other end of the die orifice.

[0028] An image acquisition device is deployed to acquire multiple video frames during the copper part extrusion process at a fixed preset frame rate, wherein the image acquisition device has a fixed posture, wherein the preset frame rate is 60 frames per second, that is, 60 image frames can be acquired per second; for any position point on the copper product, it moves at a fixed extrusion speed in a direction away from the other end of the mold orifice, and according to the relationship between the extrusion speed and the preset frame rate, the pixel coordinates of any position point on the copper product in each video frame can be located.

[0029] Specifically, intercepting the neighborhood image of any position point in each video frame according to the extrusion speed includes: determining the pixel coordinates of the position point in any video frame according to the extrusion speed, intercepting the video frame according to a preset number of pixels on both sides of the pixel coordinates, and obtaining the neighborhood image of the position point in the video frame. The preset number is 1.

[0030] Among them, when any position point enters the field of view of the image acquisition device, the neighborhood image of the position point can be acquired, until the position point moves out of the field of view of the image acquisition device, multiple neighborhood images of the position point can be acquired.

[0031] In this way, during the extrusion molding process of the copper part extruder, multiple neighborhood images of any position point on the copper product are collected, and one neighborhood image corresponds to one timestamp.

[0032] For location points , timestamp The vibration amplitude Satisfies the relationship: , , and Location points In timestamp , Timestamp and timestamp The neighborhood image of Indicates calculating the average pixel value.

[0033] Understandably, For timestamp The difference map between the neighborhood image of and the previous adjacent neighborhood image; For timestamp The difference map between the neighborhood image and the next adjacent neighborhood image; the larger the average pixel value in the difference map, the more relative motion occurs between the two adjacent neighborhood images. Therefore, the maximum value of the average pixel value in the two difference maps corresponding to a neighborhood image is taken as the vibration amplitude of the position point to accurately quantify the vibration amplitude of the position point at the timestamp corresponding to the neighborhood image.

[0034] For any position point, if the vibration amplitude of a timestamp is less than the preset amplitude, it means that the position point has not experienced a large vibration at the timestamp and is in a stable state. The timestamp is a stable moment, where the preset amplitude is 1; in this way, all the stable moments of the position point are taken as the stable moment set of the position point.

[0035] S102, initialize the initial segment of any position point, in response to the intersection of the initial segment stable moment set and the stable moment sets of the adjacent position points of the initial segment being not empty, add the adjacent position points to the initial segment until the intersection of the stable moment sets is empty, and obtain the stable segment of the position point, the initial segment stable moment set being the intersection of the stable moment sets of each position point in the initial segment.

[0036] In one embodiment, after obtaining a stable moment set of any position point on the copper product, a plurality of stable segments can be obtained according to the stable moment set of the position point, and one stable segment includes at least one stable moment.

[0037] For location points For example, the location point How to obtain the stable segment. First, initialize the position point The initial segment includes the position point , at this time, the stable time set of the initial segment is the position point The stationary moment set of the initial segment adjacent position points includes the position point The adjacent position points on both sides, including the position point and location points , here the location point For example, when the initial segment stable moment set and the initial segment adjacent position points When the intersection of the stationary moment sets is not empty, it means that the initial segment and the adjacent position points At the same time stamp, they are all at a stable moment, so the adjacent position points can be Add location point The initial segment of the realization position point Update of the initial segment, position point The initial segment includes the location point and location points .

[0038] At this time, the stable moment set of the initial segment is the position point and location points The intersection of the stationary moment sets, the initial segment adjacent position points are position points and location points The adjacent position points on both sides of the position segment, including the position points and location points In this way, the adjacent position points of the initial segment are continuously added to the initial segment until the intersection of the stable time set of the initial segment and the stable time set of the adjacent position points of the initial segment is empty, indicating that the adjacent position points of the initial segment cannot maintain a stable state at the same timestamp as the position points in the initial segment.

[0039] It can be understood that if the initial segment stable time set and the initial segment adjacent position points The intersection of the stationary time sets includes 10 timestamps, which means that all the position points in the initial segment and the adjacent position points in the initial segment At these 10 time stamps, they were all in a stable state.

[0040] In this way, multiple stable segments are obtained, one stable segment includes multiple position points, and the multiple position points in one stable segment can be in a stable state at the same time stamp.

[0041] S103, splicing the neighborhood images of each position point in the stable segment according to the stable moment set of the stable segment to obtain an extruded image segment of the stable segment.

[0042] In one embodiment, obtaining the extruded image segment of the stable segment includes: obtaining a video frame with an arbitrary time stamp in the stable moment set of the stable segment, intercepting the video frame according to the pixel point coordinates of each position point in the stable segment in the video frame, and obtaining the initial image segment of the stable segment; and taking the average of all the initial image segments as the extruded image segment of the stable segment.

[0043] The stable moment set of the stable segment includes at least one timestamp. The pixel coordinates of each position point in the stable segment in the video frame corresponding to the timestamp are obtained, and the video frame width is recorded as , the height is recorded as , after intercepting the video frame corresponding to the timestamp according to the pixel coordinates, the width is , the height is Initial image segment, where is the number of position points in the stable segment; the number of the initial image segments is the same as the number of concentrated time stamps at the stable moment of the stable segment.

[0044] In another embodiment, since one stable segment can collect initial image segments of multiple timestamps, and there are differences in the vibration amplitudes of various positions in the stable segment at different timestamps, the average of all initial image segments is directly used as the extruded image segment, and the differences in the vibration amplitudes of various positions in the stable segment at different timestamps are ignored, so that the extruded image segment is greatly affected by the vibration amplitude, and the image information of the copper part in the stable state cannot be accurately obtained. Therefore, in the process of calculating the average of all initial image segments, the vibration amplitude of each position in different initial image segments (that is, at different timestamps) can be considered.

[0045] Specifically, the extruding image segment by taking the mean of all the initial image segments as the stable segment comprises: The pixel value of any position point in each initial image segment is obtained to obtain a pixel value sequence; the weighting coefficient of each initial image segment is determined according to the vibration amplitude of the position point at the corresponding time stamp of the initial image segment, and the weighting coefficient is negatively correlated with the vibration amplitude; the pixel value sequence is weighted averaged according to the weighting coefficient to obtain the weighted pixel value of the position point, and the weighted pixel values ​​of all position points in the stable segment constitute the extruded image segment.

[0046] Among them, the location point In timestamp The weighting factor for: , For location point In timestamp The vibration amplitude, For location point The sum of the vibration amplitudes at corresponding timestamps in all initial image segments.

[0047] In this way, the extrusion image segments of each stable segment are obtained. The extrusion image segments are image information of the copper part in a stable state, which can eliminate the noise caused by vibration in the video frame, thereby improving the accuracy of the extrusion molding detection result.

[0048] S104, splicing the extruded image segments to obtain an extruded image, and obtaining a detection result based on the extruded image.

[0049] In one embodiment, an extruded image segment includes pixel values ​​of multiple positions in a steady state. After all the extruded image segments are spliced, an extruded image can be obtained, and the extruded image includes pixel values ​​of all positions on the copper part in a steady state.

[0050] Specifically, the stitching of the extruded image segments includes: calculating a difference map of the overlapping areas of any two extruded image segments, updating the translation parameters of each extruded image segment, and completing the stitching of the two extruded image segments when the average pixel value of the difference map reaches a minimum value.

[0051] Among them, the overlapping area of ​​the two extruded image segments can be accurately located according to the position points contained in the extruded image segments. If the area of ​​the overlapping area is equal to 0, it means that the two extruded image segments cannot be spliced ​​together. Otherwise, the two extruded image segments can be spliced ​​together according to the overlapping area until a complete extruded image is obtained.

[0052] Since the overlapping area has been determined, the translation parameters only include up and down translation parameters.

[0053] It is understandable that if any position point is always in a state of large vibration amplitude during the copper extrusion process, there will be no overlapping area between the two extrusion image segments, and splicing will not be possible, which means that the complete extrusion image cannot be obtained. At this time, it means that the copper extruder is in a continuous and relatively violent vibration, the copper extruder is in an abnormal working state, and an early warning message of abnormal operation status of the copper extruder is issued.

[0054] In one embodiment, obtaining the detection result based on the extrusion image includes: dividing the extrusion image into multiple image blocks along the extrusion direction of the copper part, calculating the similarity between any image blocks to construct a similarity matrix; calculating the Euclidean distance between the similarity matrix and the unit matrix, and in response to the Euclidean distance being greater than a preset threshold, the detection result is abnormal quality, otherwise, the detection result is normal quality.

[0055] The unit matrix is ​​a matrix with the same size as the similarity matrix and all values ​​are 1. The size of the similarity matrix is ​​related to the number of image blocks. If the number of image blocks is 5, the size of the similarity matrix is ​​a square matrix with 5 rows and 5 columns.

[0056] The preset threshold is 0.5. It can be understood that in response to the Euclidean distance being greater than the preset threshold, it indicates that there are large differences in the image features of the copper parts between the image blocks, and defects such as uneven thickness, surface cracks, or even fractures appear on the surface of the copper parts. At this time, the detection result is abnormal quality, and quality detection of the copper parts during extrusion is achieved.

[0057] It should be noted that, for ordinary technicians in this field, several modifications and improvements can be made without departing from the concept of this application, which all belong to the protection scope of this application. Therefore, the protection scope of the patent of this application shall be based on the attached claims.

Claims

1. A copper extrusion machine extrusion molding detection method, characterized in that: The detection method comprises: Collect multiple video frames of the copper part extrusion process, and intercept the neighborhood image of any position point in each video frame according to the extrusion speed; calculate the difference map between any neighborhood image and two adjacent neighborhood images, and use the maximum value of the average pixel value in the difference map as the vibration amplitude of the position point at the corresponding timestamp of the neighborhood image, and use the timestamp with a vibration amplitude less than a preset amplitude as the stable moment, so as to obtain the stable moment set of each position point; Initialize an initial segment of an arbitrary position point, and in response to the intersection of the initial segment stable time set and the stable time set of the position points adjacent to the initial segment being not empty, add the adjacent position points to the initial segment until the intersection of the stable time sets is empty, thereby obtaining the stable segment of the position point, wherein the initial segment stable time set is the intersection of the stable time sets of each position point in the initial segment; According to the stable moment set of the stable segment, the neighborhood images of each position point in the stable segment are spliced ​​to obtain the extruded image segment of the stable segment; The extruded image segments are spliced ​​to obtain an extruded image, and the detection result is obtained according to the extruded image.

2. A copper extruder extrusion molding detection method according to claim 1, characterized in that: The method of intercepting the neighborhood image of any position point in each video frame according to the extrusion speed includes: The pixel coordinates of the position point in any video frame are determined according to the extrusion speed, and the video frame is intercepted according to a preset number of pixels on both sides of the pixel coordinates to obtain a neighborhood image of the position point in the video frame.

3. A copper extruder extrusion molding detection method according to claim 1, characterized in that: Location Point Timestamp The vibration amplitude for: , , and Position points In timestamp , Timestamp and timestamp The neighborhood image of Indicates calculating the average pixel value.

4. A copper extruder extrusion molding detection method according to claim 1, characterized in that: The initializing the initial segment of the arbitrary position point includes: the initial segment includes the arbitrary position point.

5. A copper extruder extrusion molding detection method according to claim 1, characterized in that: The extruded image segments that result in a stable segment include: Obtain a video frame with an arbitrary time stamp in the stable moment set of the stable segment, intercept the video frame according to the pixel point coordinates of each position point in the stable segment in the video frame, and obtain an initial image segment of the stable segment; The mean of all initial image segments is taken as the extruded image segment of the stationary segment.

6. A copper extruder extrusion molding detection method according to claim 5, characterized in that: The extruded image segment taking the mean of all the initial image segments as the stable segment comprises: Obtain the pixel value of any position point in each initial image segment to obtain a pixel value sequence; Determining a weighting coefficient of each initial image segment according to the vibration amplitude of the position point at the time stamp corresponding to the initial image segment, wherein the weighting coefficient is negatively correlated with the vibration amplitude; The pixel value sequence is weighted averaged according to the weighting coefficient to obtain the weighted pixel value of the position point, and the weighted pixel values ​​of all the position points in the stable segment constitute the extruded image segment.

7. A copper extruder extrusion molding detection method according to claim 6, characterized in that: Location Point In timestamp The weighting factor for: , For location point In timestamp The vibration amplitude, For location point The sum of the vibration amplitudes at corresponding timestamps in all initial image segments.

8. A copper extruder extrusion molding detection method according to claim 1, characterized in that: The splicing of the extruded image segments comprises: The difference map of the overlapping area of ​​any two extruded image segments is calculated, and the translation parameters of each extruded image segment are updated until the average pixel value of the difference map reaches the minimum value, and the splicing of the two extruded image segments is completed.

9. A copper extruder extrusion molding detection method according to claim 1, characterized in that: The detection results obtained based on the extrusion image include: The extrusion image is equally divided into multiple image blocks along the extrusion direction of the copper part, and the similarity between any image blocks is calculated to construct a similarity matrix; The Euclidean distance between the similarity matrix and the unit matrix is ​​calculated. In response to the Euclidean distance being greater than a preset threshold, the detection result is abnormal quality, otherwise, the detection result is normal quality.

10. A copper extrusion machine extrusion molding detection method according to claim 1, characterized in that: In response to the failure to obtain the extrusion image, an early warning message indicating that the copper extrusion machine is operating abnormally is issued.

Images