Visual monitoring method for stacking production of fan-shaped generators based on image processing

Abstract

The present invention relates to the field of image edge detection technology, and in particular to a method for visual monitoring of the stacking production of fan-shaped blade generators based on image processing. The method determines an edge deformation coefficient corresponding to each edge line segment according to the extension performance of the extruded edge under the influence of extrusion stress; determines an extrusion depression coefficient corresponding to each edge line segment based on the grayscale performance presented by the extrusion depression; thereby comprehensively representing the extrusion edge possibility of each edge line segment through the edge deformation coefficient and the extrusion depression coefficient, so that the extrusion edge of the motor fan-shaped blade represented by the real extrusion edge screened out according to the extrusion edge possibility is more accurate, so that the effect of visual monitoring of the stacking production of fan-shaped blade generators is better.

Description

Technical Field

[0001] The invention relates to the technical field of image edge detection, and in particular to a visual monitoring method for the lamination production of fan-shaped generators based on image processing. Background Art

[0002] Segment lamination generally refers to a process used in manufacturing the rotor of a rotating motor or generator. In this process, thin metal sheets of sector or similar shapes, such as silicon steel sheets with good magnetic properties and the most widely used, are stacked together to form the magnetic core of the rotor or stator. The segment lamination process may cause improper alignment, resulting in deformation or damage, which affects the quality of the magnetic circuit. Therefore, it is necessary to monitor the generator segments after lamination to timely screen out the scrapped segments that are deformed or damaged, thereby improving the quality of the magnetic circuit.

[0003] Considering that the efficiency and accuracy of manual visual inspection of the fan-shaped surface are low, the surface image of the fan-shaped piece after lamination is usually analyzed and processed based on the image processing method to further detect the scrapped fan-shaped piece. Considering that when the fan-shaped piece after lamination is damaged, there will be a large number of edges formed by extrusion on the corresponding image; therefore, based on this feature, the existing technology usually measures the damage of the generator fan-shaped piece based on the number of edge line segments on the edge image corresponding to the fan-shaped piece based on the edge detection method, so as to perform visual monitoring of the fan-shaped piece generator lamination production.

[0004] However, in the fan-shaped sheet stacking environment, stains and other impurities may appear on the surface of the fan-shaped sheet. The texture of these stains or impurities will usually appear as edge segments on the edge image corresponding to the fan-shaped sheet, resulting in that the edge segments on the edge image corresponding to the fan-shaped sheet are not all real extruded edges. Therefore, the existing technology will mistake the texture corresponding to the stains or impurities on the fan-shaped sheet surface as the texture of the extruded edge of the motor fan-shaped sheet, resulting in poor accuracy in measuring the damage status of the generator fan-shaped sheet according to the number of edge segments on the edge image corresponding to the fan-shaped sheet, thereby making the visual monitoring of the fan-shaped sheet generator stacking production less effective. Summary of the invention

[0005] The present application provides a method for visual monitoring of the stacking production of fan-shaped blade generators based on image processing. The method determines the edge deformation coefficient corresponding to each edge segment based on the extension performance of the extruded edge under the influence of extrusion stress; determines the extrusion depression coefficient corresponding to each edge segment based on the grayscale performance of the extrusion depression; thereby comprehensively representing the extrusion edge possibility of each edge segment through the edge deformation coefficient and the extrusion depression coefficient, so that the extrusion edge of the motor fan-shaped blade represented by the real extrusion edge screened out according to the extrusion edge possibility is more accurate, and solves the problem that the prior art will mistake the texture corresponding to the stain or impurity on the surface of the fan-shaped blade as the texture of the extruded edge of the motor fan-shaped blade, resulting in poor accuracy in measuring the damage condition of the generator fan-shaped blade according to the number of edge segments on the edge image corresponding to the fan-shaped blade, so that the effect of visual monitoring of the stacking production of fan-shaped blade generators is better.

[0006] In a first aspect, the present application provides a method for visually monitoring the lamination production of a fan-shaped generator based on image processing, comprising:

[0007] Perform edge detection based on the grayscale image of the surface of the generator fan-shaped pieces after the fan-shaped pieces are stacked, and determine the fan-shaped piece edges and all edge segments;

[0008] Based on the influence of the extrusion stress of the fan-shaped sheet, the edge deformation coefficient of each edge line segment is determined according to the extension direction of each edge line segment, the straight line similarity and the distance between each edge line segment and the edge of the fan-shaped sheet;

[0009] On the grayscale image of the surface of the sector sheet, based on the grayscale representation of the extrusion depression, and according to the grayscale change trend and the gradient change trend within the neighborhood of each edge segment, the extrusion depression coefficient of each edge segment is determined;

[0010] Determine the extrusion edge possibility of each edge segment based on the edge deformation coefficient and the extrusion concave coefficient; and screen out the real extrusion edges in all edge segments according to the extrusion edge possibility;

[0011] Visual monitoring of the lamination production of fan-shaped generators is performed based on the actual extrusion edge.

[0012] In an embodiment of the first aspect of the present application, the method for obtaining the edge deformation coefficient includes:

[0013] Based on the shortest distance from each edge pixel point on each edge line segment to the edge of the fan-shaped piece, determine the edge distance of each edge pixel point;

[0014] According to the number of edge pixels, the edge distance range and the minimum edge distance on each edge segment, the position direction squeezing index of each edge segment is determined; the edge distance range is positively correlated with the position direction squeezing index, and the number of edge pixels and the minimum edge distance are negatively correlated with the position direction squeezing index;

[0015] Determine the edge smoothness corresponding to each edge line segment, which represents the smoothness of the edge contour, and the straight line similarity corresponding to each edge line segment, which represents the similarity of the fitted straight lines;

[0016] Based on the position direction extrusion index, the edge smoothness and the straight line similarity, the edge deformation coefficient of each edge segment is determined; the position direction extrusion index, the edge smoothness and the straight line similarity are all positively correlated with the edge deformation coefficient.

[0017] In an embodiment of the first aspect of the present application, the method for obtaining the position direction squeezing index includes:

[0018] Determine an edge pointing coefficient of each edge line segment according to a ratio of the edge distance extreme difference to the number of edge pixels;

[0019] The position direction extrusion index of each edge line segment is determined by multiplying the negative correlation mapping value of the minimum edge distance value by the edge orientation coefficient.

[0020] In an embodiment of the first aspect of the present application, the method for obtaining the edge smoothness includes:

[0021] Based on the chain code method, the chain code corresponding to each edge segment is counted; on the chain code corresponding to each edge segment, the difference between each chain code value and the next adjacent chain code value is used as the local chain code change value of each chain code value;

[0022] The edge smoothness of each edge segment is determined according to the negative correlation mapping value of the mean of the local chain code change values ​​of all chain code values ​​on the chain code corresponding to each edge segment.

[0023] In an embodiment of the first aspect of the present application, the method for obtaining the degree of straight line similarity includes:

[0024] Determine the fitting straight line corresponding to each edge segment based on a straight line fitting method;

[0025] Perform dynamic time warping on each edge segment and the corresponding fitting line to determine the DTW distance between each edge segment and the corresponding fitting line;

[0026] The DTW distance is subjected to negative correlation mapping processing to determine the straight line similarity of each edge line segment.

[0027] In an embodiment of the first aspect of the present application, the method for obtaining the extrusion depression coefficient includes:

[0028] On each edge line segment, a preset number of neighboring pixel points that are closest to each edge pixel point in a preset first reference direction are arranged according to the preset first reference direction to determine a first adjacent pixel point sequence for each edge pixel point; on each edge line segment, a preset number of neighboring pixel points that are closest to each edge pixel point in a preset second reference direction are arranged according to the preset second reference direction to determine a second adjacent pixel point sequence for each edge pixel point; wherein both the preset first reference direction and the preset second reference direction are perpendicular to the fitting straight line of the edge line segment, and the preset first reference direction and the preset second reference direction are opposite;

[0029] Determine the grayscale characteristic value of each pixel point according to the grayscale value and grayscale gradient value of each pixel point; the grayscale value and grayscale gradient value are both positively correlated with the grayscale characteristic value;

[0030] On each edge line segment, the degree of extrusion depression of each edge pixel point is determined according to the grayscale characteristic value distribution deviation between the first adjacent side pixel point sequence and the second adjacent side pixel point sequence corresponding to each edge pixel point;

[0031] The average of the extrusion and depression degrees of all edge pixels on each edge segment is normalized to obtain the extrusion and depression coefficient of each edge segment.

[0032] In an embodiment of the first aspect of the present application, the method for obtaining the degree of extrusion depression includes:

[0033] sequentially taking the first adjacent side pixel point sequence and the second adjacent side pixel point sequence as a target pixel point sequence;

[0034] Determine the grayscale feature change value of each pixel in the target pixel sequence according to the grayscale feature value difference between each pixel and the previous pixel in the target pixel sequence;

[0035] The grayscale feature change values ​​of all pixels in the target pixel sequence are averaged to obtain the overall change feature value of the target pixel sequence;

[0036] The degree of extrusion and depression of each edge pixel point on each edge line segment is determined according to the difference between the overall change characteristic value of the first adjacent side pixel point sequence and the overall change characteristic value of the second adjacent side pixel point sequence.

[0037] In an embodiment of the first aspect of the present application, the method for obtaining the possibility of extruding an edge includes:

[0038] The product of the edge deformation coefficient and the extrusion concave coefficient is normalized to determine the extrusion edge possibility of each edge segment.

[0039] In an embodiment of the first aspect of the present application, the method for obtaining the real extruded edge includes:

[0040] The edge line segment whose corresponding extrusion edge possibility is greater than the preset edge threshold is regarded as the real extrusion edge.

[0041] In an embodiment of the first aspect of the present application, the method for visually monitoring the lamination production of fan-shaped generators according to the real extrusion edge includes:

[0042] When the number of real extruded edges in the grayscale image of the surface of the fan-shaped piece is greater than or equal to a preset scrap threshold, the corresponding generator fan-shaped piece is determined to be a scrapped fan-shaped piece;

[0043] When the number of real extruded edges in the grayscale image of the surface of the sector piece is less than a preset scrap threshold, the corresponding generator sector piece is determined to be a normal sector piece.

[0044] In a second aspect, the present application provides a visual monitoring system for fan-shaped generator stacking production based on image processing, the system comprising:

[0045] A data acquisition preprocessing module is used to perform edge detection based on the grayscale image of the surface of the generator fan-shaped pieces after the fan-shaped pieces are stacked, and determine the fan-shaped piece edges and all edge segments;

[0046] A first determination module is used to determine the edge deformation coefficient of each edge line segment based on the influence of the extrusion stress of the sector piece and according to the extension direction of each edge line segment, the straight line similarity and the distance between each edge line segment and the edge of the sector piece;

[0047] A second determination module is used to determine the extrusion depression coefficient of each edge line segment on the grayscale image of the surface of the sector-shaped piece based on the grayscale expression of the extrusion depression and the grayscale change trend and gradient change trend within the neighborhood of each edge line segment;

[0048] The fan-shaped blade generator stacking production visual monitoring module is used to determine the extrusion edge possibility of each edge segment based on the edge deformation coefficient and the extrusion depression coefficient; screen out the real extrusion edges in all edge segments according to the extrusion edge possibility; and perform visual monitoring of the fan-shaped blade generator stacking production according to the real extrusion edges.

[0049] In a third aspect, the present application provides a computer device, comprising a memory and a processor. The memory is used to store computer program code, and the processor is used to call and run the computer program code from the memory to execute the method of the first aspect or any embodiment of the first aspect of the present application.

[0050] In a fourth aspect, the present application provides a computer program product, comprising a computer program code, which, when executed, performs the method of the first aspect of the present application or any embodiment of the first aspect.

[0051] In a fifth aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program code. When the computer program code is executed, the method of the first aspect or any embodiment of the first aspect of the present application is performed.

[0052] This application has the following beneficial effects:

[0053] The reason why the existing technology has poor accuracy in measuring the damage of the generator fan blades based on the number of edge segments on the edge image corresponding to the fan blades is that the texture of stains or impurities on the surface of the fan blades is usually also displayed as edge segments on the edge image corresponding to the fan blades, causing the texture corresponding to the stains or impurities on the surface of the fan blades to be mistaken for the texture of the extruded edge of the motor fan blades. Therefore, in order to improve the accuracy of measuring the damage of the generator fan blades, it is necessary to eliminate or reduce the influence of stains or impurities on the surface of the fan blades.

[0054] On the fan-shaped sheet, the presence of stains and impurities is random, and the generation of the real extrusion edge is affected by the extrusion during the fan-shaped sheet stacking process. Extrusion will cause the fan-shaped sheet surface to have certain deformation characteristics. Therefore, the real extrusion edge has certain distinguishing characteristics relative to stains and impurities. The possibility of the extrusion edge can be calculated based on the distinguishing characteristics, so as to screen out a more accurate real extrusion edge, so as to eliminate or reduce the influence of stains or impurities on the surface of the fan-shaped sheet. After the fan-shaped sheet is stacked, the corresponding real extrusion edge will be formed at the edge or boundary area of ​​the fan-shaped sheet under the influence of the extrusion stress, and usually extends in a straight line perpendicular to the edge of the fan-shaped sheet. Therefore, the edge deformation coefficient of each edge segment is determined in combination with this feature, that is, the possibility of belonging to the real extrusion edge in the dimension of morphological position. During the lamination process, the placement of the fan-shaped blades and other reasons will lead to uneven force, and the extrusion stress will diffuse outward, resulting in strip-shaped extrusion depressions at the edge. Therefore, the extrusion depression coefficient of each edge segment is determined according to the grayscale gradient distribution and grayscale distribution of the depression area, that is, the possibility of belonging to a real extrusion edge in the depression feature dimension; the edge deformation coefficient and the extrusion depression coefficient are further combined to comprehensively characterize the possibility of the extrusion edge of each edge segment, so as to screen out more accurate real extrusion edges, minimize the influence of stains or impurities, and make the measurement of the damage of the generator fan-shaped blades based on the real extrusion edges more accurate, that is, the visual monitoring of the fan-shaped blade generator lamination production based on the real extrusion edges is more effective. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0056] Figure 1 A flow chart of a method for visual monitoring of fan-shaped generator stacking production based on image processing provided by one embodiment of the present invention.

[0057] Figure 2 A fan-shaped slice edge image based on a canny edge detection algorithm provided by an embodiment of the present invention;

[0058] Figure 3 A flow chart of a method for obtaining edge deformation coefficients of a visual monitoring method for lamination production of fan-shaped generators based on image processing provided by one embodiment of the present invention;

[0059] Figure 4A flow chart of a method for obtaining an extrusion depression coefficient of a visual monitoring method for lamination production of a fan-shaped generator based on image processing provided by one embodiment of the present invention;

[0060] Figure 5 A structural diagram of a visual monitoring system for stacking production of fan-shaped generators based on image processing provided by one embodiment of the present invention;

[0061] Figure 6 A schematic diagram of the structure of a computer device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0062] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation method, structure, features and effects of a visual monitoring method for the stacking production of fan-shaped generators based on image processing proposed by the present invention, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment, and the specific features, structures or characteristics in one or more embodiments may be combined in any suitable form. In addition, the terms "first" and "second" are used only for descriptive purposes and cannot be understood as implying or suggesting relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features.

[0063] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0064] The specific scheme of the visual monitoring method for the stacking production of fan-shaped generators based on image processing provided by the present invention is described in detail below with reference to the accompanying drawings.

[0065] The present application embodiment provides a visual monitoring method for the stacking production of fan-shaped generators based on image processing. Figure 1 , which shows a flow chart of a method for visual monitoring of the lamination production of fan-shaped generators based on image processing provided by an embodiment of the present invention, the method comprising:

[0066] Step S101: performing edge detection based on the grayscale image of the surface of the generator sector pieces after the sector pieces are stacked, and determining the sector piece edges and all edge segments.

[0067] In the automatic production process of fan-shaped sheet lamination, a mechanical arm is used to stack the fan-shaped sheets, and the stacking is performed once. An industrial camera is installed above the mechanical arm. After the lamination is completed, the surface of the fan-shaped sheet is photographed by the industrial camera to obtain the initial image of the surface of the fan-shaped sheet for image processing. Considering that the grayscale feature needs to be combined for subsequent analysis, the initial image of the surface of the fan-shaped sheet is grayed to obtain a grayscale image of the surface of the fan-shaped sheet; the implementer can adjust the image acquisition device according to the specific implementation environment, and no further details are given here.

[0068] In a specific implementation method of an embodiment of the present invention, the initial image of the fan-shaped piece surface is grayed and then subjected to mean filtering processing to obtain a grayscale image of the fan-shaped piece surface. The implementer can adjust the filtering method according to the specific implementation environment, such as median filtering, etc. De-noising through filtering can minimize the impact of noise on the grayscale image of the fan-shaped piece surface, so that the real extrusion edge obtained by subsequent analysis is more accurate.

[0069] Considering that the real extruded edge is an edge line segment, in order to extract the edge line segment in the grayscale image of the fan-shaped piece surface, it is necessary to perform edge detection on the grayscale image of the fan-shaped piece surface. In a specific implementation of an embodiment of the present invention, the edge detection method adopts the canny edge detection algorithm. Due to the influence of the extrusion stress during the overlapping process, the real extruded edge is usually close to the edge of the fan-shaped piece. Therefore, it is necessary to obtain the edge of the fan-shaped piece in the fan-shaped piece edge image after the edge detection of the fan-shaped piece grayscale image to facilitate subsequent analysis. In a specific implementation of an embodiment of the present invention, the edge corresponding to the connected domain with the largest edge connected domain area in the fan-shaped piece edge image is taken as the edge of the fan-shaped piece. Please refer to Figure 2 , which shows a fan-shaped slice edge image based on a canny edge detection algorithm provided by an embodiment of the present invention; Figure 2 In the example, the edge of the fan-shaped piece is the edge corresponding to the largest circular contour, and Figure 2 There are a large number of unclosed edge segments, which are the edge segments required by the embodiment of the present invention. It should be noted that the Canny edge detection method is a prior art well known to those skilled in the art, and will not be further defined or elaborated herein.

[0070] Step S102: Based on the influence of the extrusion stress of the sector piece, the edge deformation coefficient of each edge line segment is determined according to the extension direction of each edge line segment, the straight line similarity and the distance between the edge of the sector piece and the edge of the sector piece.

[0071] On the fan-shaped piece, the existence of stains and impurities is random, and the generation of the real extruded edge is affected by the extrusion during the stacking process of the fan-shaped piece. The extrusion will cause the surface of the fan-shaped piece to have certain deformation characteristics. Therefore, the real extruded edge has certain distinguishing characteristics relative to the stains and impurities. The possibility of the extruded edge can be calculated based on the distinguishing characteristics, so as to screen out a more accurate real extruded edge, so as to achieve the purpose of eliminating or reducing the influence of stains or impurities on the surface of the fan-shaped piece.

[0072] After the fan-shaped pieces are stacked, the corresponding real extrusion edge will be formed at the edge of the fan-shaped piece or the boundary area under the influence of the extrusion stress, and usually extends in a straight line perpendicular to the edge of the fan-shaped piece. That is, the real extrusion edge usually points to the edge of the fan-shaped piece on the basis of approximating a straight line, and the distance between the edge of the fan-shaped piece is close. Therefore, the edge deformation coefficient of each edge segment is determined in three dimensions: extension direction, straight line similarity, and distance from the edge of the fan-shaped piece, so as to characterize the possibility that the edge segment belongs to the real extrusion edge in terms of morphological position characteristics.

[0073] In some possible implementations of the present invention, the specific implementation process of the method for obtaining the edge deformation coefficient in step S102 includes: Figure 3 , which shows a flow chart of a method for obtaining edge deformation coefficient of a visual monitoring method for stacking production of fan-shaped generators based on image processing provided by an embodiment of the present invention, the method comprising:

[0074] Step S301: determining the edge distance of each edge pixel point based on the shortest distance from each edge pixel point on each edge line segment to the edge of the fan-shaped slice.

[0075] The real extruded edge is usually closer to the edge of the fan-shaped piece, so the edge distance of each edge pixel is calculated first to facilitate the subsequent measurement of the possibility of each edge segment belonging to the real extruded edge in the position dimension.

[0076] Step S302: determining the position direction extrusion index of each edge line segment according to the number of edge pixels, the edge distance extreme difference and the edge distance minimum value on each edge line segment.

[0077] Since the edge distance is the shortest distance from each edge pixel to the edge of the fan-shaped piece, the larger the edge distance range of the edge segment, the more likely the extension direction of the edge segment is to point to the edge of the fan-shaped piece, that is, the greater the possibility of being a real extruded edge in the direction dimension; however, considering that the number of edge pixels will affect the size of the edge distance range, the edge distance range is limited by the number of edge pixels; that is, the larger the edge distance range and the smaller the number of pixels, the more likely the corresponding edge segment is a real extruded edge. The minimum edge distance measures the distance between each edge segment and the edge of the fan-shaped piece. The smaller the minimum edge distance, the closer the edge segment is to the edge of the fan-shaped piece, and the greater the possibility of being a real extruded edge. Therefore, if the position direction extrusion index is used to characterize the possibility that the edge segment is a real extruded edge in the position and extension direction dimensions, the edge distance range is positively correlated with the position direction extrusion index, and the number of edge pixels and the minimum edge distance are negatively correlated with the position direction extrusion index.

[0078] Preferably, in some possible implementations of the embodiments of the present invention, the method for obtaining the position direction squeezing index in step S302 includes:

[0079] According to the ratio of the edge distance range to the number of edge pixels, the edge direction coefficient of each edge segment is determined; the position direction extrusion index of each edge segment is determined by multiplying the negative correlation mapping value of the minimum edge distance with the edge direction coefficient. Using the number of edge pixels as the denominator of the edge distance range reduces the influence of the number of edge pixels on the extension direction measurement. The minimum edge distance is negatively correlated with the position direction extrusion index, so the negative correlation mapping value of the minimum edge distance and the edge direction coefficient are combined in the form of a product.

[0080] In a specific implementation of the embodiment of the present invention, the process of obtaining the position direction squeezing index is expressed by the formula: ;in, For the The position direction extrusion index of the edge segment; For the The maximum value of the edge distance corresponding to all edge pixels on the edge segment; For the The minimum edge distance corresponding to all edge pixels on the edge segment; For the The number of edge pixels on an edge segment; is an exponential function with a natural constant as base; is the absolute value symbol.

[0081] Step S303: determining the edge smoothness corresponding to each edge line segment, which represents the smoothness of the edge contour, and the line similarity corresponding to each edge line segment, which represents the similarity of the fitted lines.

[0082] Since the real extruded edge usually extends in a straight line perpendicular to the edge of the fan-shaped piece, the direction of the edge segment corresponding to the real extruded edge changes relatively steadily and is relatively smooth overall; and the overall approximation is that the straight line points to the fan-shaped edge. Therefore, the greater the edge smoothness corresponding to each edge segment, which represents the smoothness of the edge contour, and the greater the straight line similarity representing the similarity of the fitted straight lines, the more likely the corresponding edge segment is to be a real extruded edge.

[0083] Preferably, in some possible implementations of the embodiments of the present invention, the method for obtaining the edge smoothness in step S303 includes:

[0084] Based on the chain code method, the chain code corresponding to each edge segment is counted; on the chain code corresponding to each edge segment, the difference between each chain code value and the next adjacent chain code value is used as the local chain code change value of each chain code value; according to the negative correlation mapping value of the mean of the local chain code change values ​​of all chain code values ​​on the chain code corresponding to each edge segment, the edge smoothness of each edge segment is determined. In a specific implementation process of an embodiment of the present invention, the chain code adopts an eight-connected chain code, and the implementer may also adopt a four-connected chain code according to the specific implementation environment, which will not be further described here.

[0085] The chain code mainly represents the direction information of the edge. For a standard straight line, the corresponding extension direction is consistent as a whole, and the corresponding chain codes are all equal, that is, the difference between all adjacent chain codes is 0; and the more curved or irregular the corresponding edge is, the stronger the volatility of the corresponding chain code sequence is, and the greater the difference between adjacent chain codes is; therefore, the local chain code change value can represent the degree of change in the edge extension direction at the position corresponding to each chain code value; therefore, the larger the overall local chain code change value of all chain code values ​​on the chain code corresponding to each edge segment, the greater the difference between the corresponding edge segment and the straight line, that is, the more irregular, the smaller the corresponding edge smoothness is. Therefore, the mean of the local chain code change value of all chain code values ​​is used for negative correlation mapping processing, and the corresponding possibility of representing the edge segment as an extruded edge in the dimension of direction change is obtained.

[0086] In a specific implementation of the embodiment of the present invention, the process of obtaining the local chain code change value is expressed by the formula: ;in, For the The chain code corresponding to the edge segment The local chain code change value of the chain code value; For the The chain code corresponding to the edge segment The value of the chain code value; For the The chain code corresponding to the edge segment The value of the chain code value; For the The chain code corresponding to the edge segment The numeric value of the chain code value.

[0087] It should be noted that for the first chain code value of the edge segment, since there is no previous chain code value, it is impossible to calculate the local chain code change value for it. Therefore, for the completeness of subsequent calculations, the local chain code change value of the second chain code value is used as the local chain code change value of the first chain code value; similarly, for the last chain code value of the edge segment, the chain code change value of the second-to-last chain code value is used as the local chain code change value of the last chain code value. In addition, it should be noted that since the edge segment is a unidirectional segment, the starting point or end point of the corresponding chain code can be randomly selected at the two endpoints of the edge segment, which will not be further described here.

[0088] Preferably, in some possible implementations of the embodiments of the present invention, the method for obtaining the straight line similarity in step S303 includes:

[0089] Based on the straight line fitting method, the fitting straight line corresponding to each edge segment is determined; each edge segment and the corresponding fitting straight line are subjected to dynamic time warping processing to determine the DTW distance between each edge segment and the corresponding fitting straight line; the DTW distance is subjected to negative correlation mapping processing to determine the straight line similarity of each edge segment.

[0090] In a specific implementation of an embodiment of the present invention, a linear fit is performed on each edge segment by the least squares method to obtain a corresponding fitted straight line. According to the principle of the Dynamic Time Warping (DTW) algorithm, the larger the DTW distance between the edge segment and the corresponding fitted straight line, the lower the similarity between the edge segment and the corresponding fitted straight line, that is, the smaller the similarity of the straight lines. Therefore, a negative correlation mapping process is performed using the DTW distance to obtain the possibility of characterizing the edge segment as an extruded edge in the dimension of straight line similarity.

[0091] In another specific implementation of the embodiment of the present invention, after performing straight line fitting on each edge segment by the least squares method to obtain the corresponding fitting straight line, the residuals obtained by the least squares method are accumulated and negatively correlated to obtain the straight line similarity of each edge segment.

[0092] Step S304: determining the edge deformation coefficient of each edge segment based on the position direction extrusion index, edge smoothness, and straight line similarity.

[0093] The position and direction extrusion index represents the possibility that the edge segment is a real extruded edge in the dimension close to the edge of the fan-shaped piece; the edge smoothness represents the possibility that the edge segment is a real extruded edge in the dimension of direction change; and the straight line similarity represents the possibility that the edge segment is a real extruded edge in the dimension of straight line similarity. Therefore, the position and direction extrusion index, edge smoothness, and straight line similarity are all positively correlated with the edge deformation coefficient.

[0094] In a specific implementation of the embodiment of the present invention, the product of the position direction extrusion index, the edge smoothness and the straight line similarity of each edge segment is used as the edge deformation coefficient of each edge segment.

[0095] In a specific implementation of the embodiment of the present invention, the process of obtaining the edge deformation coefficient is expressed by the formula: ;in, For the The edge deformation coefficient of the edge line segment; For the The position direction extrusion index of the edge segment; For the The edge smoothness of each edge segment; For the The degree of straight line similarity of edge segments; For the The number of chain code values ​​corresponding to the edge segments; For the The chain code corresponding to the edge segment The local chain code change value of the chain code value; For the The DTW distance between an edge segment and the corresponding fitting line. is an exponential function with a natural constant as its base.

[0096] Step S103: on the grayscale image of the surface of the sector sheet, based on the grayscale representation of the extrusion depression, and according to the grayscale variation trend and gradient variation trend within the neighborhood of each edge segment, determine the extrusion depression coefficient of each edge segment.

[0097] During the lamination process, the placement of the fan-shaped blades and other reasons will lead to uneven force, and the extrusion stress will spread outward, resulting in strip-shaped extrusion depressions at the edge. Therefore, the extrusion depression coefficient of each edge segment is determined according to the grayscale gradient distribution and grayscale distribution of the depression area, that is, the possibility of belonging to the real extrusion edge in the depression feature dimension. The real extrusion edge is usually located at the edge of the depression area, that is, the junction of the depression area and the normal area. In the grayscale image of the fan-shaped sheet surface, according to the grayscale performance of the depression area, in the order from the outside to the inside of the depression area, the corresponding grayscale value gradually decreases and the grayscale gradient also gradually decreases; while the grayscale value and grayscale gradient of the normal area are relatively stable; therefore, there are differences in the grayscale change trend and gradient change trend on both sides of the real extrusion edge at the junction of the depression area and the normal area. According to this feature, the extrusion depression coefficient corresponding to each edge segment that characterizes the possibility of the real extrusion edge in the neighborhood grayscale performance dimension is calculated.

[0098] Preferably, in some possible implementations of the embodiments of the present invention, the specific implementation process of the method for obtaining the extrusion depression coefficient in step S103 includes: Figure 4 , which shows a flow chart of a method for obtaining an extrusion depression coefficient of a visual monitoring method for stacking production of a fan-shaped generator based on image processing provided by an embodiment of the present invention, the method comprising:

[0099] Step S401: determining a first adjacent pixel sequence and a second adjacent pixel sequence for each edge pixel.

[0100] On each edge line segment, the preset number of neighboring pixels closest to each edge pixel point in the preset first reference direction are arranged according to the preset first reference direction to determine the first adjacent pixel sequence of each edge pixel point; on each edge line segment, the preset number of neighboring pixels closest to each edge pixel point in the preset second reference direction are arranged according to the preset second reference direction to determine the second adjacent pixel sequence of each edge pixel point; wherein the preset first reference direction and the preset second reference direction are both perpendicular to the fitting straight line of the edge line segment, and the preset first reference direction and the preset second reference direction are opposite; that is, the pixel sequence on both sides of each edge pixel point is obtained, the first adjacent pixel sequence is the pixel sequence on one side, and the second adjacent pixel sequence is the pixel sequence on the other side. In a specific implementation of an embodiment of the present invention, the preset number of neighbors is set to 10, and the implementer can adjust it according to the specific implementation environment.

[0101] Step S402: Determine the grayscale characteristic value of each pixel point according to the grayscale value and grayscale gradient value of each pixel point.

[0102] Since the grayscale value and grayscale gradient of the concave area gradually decrease from the outside to the inside, the change rules of the grayscale value and the grayscale gradient tend to be consistent from the outside to the inside, so the grayscale value and the grayscale gradient value are set to be positively correlated with the grayscale eigenvalue; by calculating the grayscale eigenvalue, the amount of calculation for subsequent calculation of the degree of extrusion depression can be reduced without changing the change relationship. In other possible implementations of the embodiments of the present invention, the grayscale value and the grayscale gradient value can also be set to be negatively correlated with the grayscale eigenvalue, as long as the relationship between the grayscale value and the grayscale gradient value and the grayscale eigenvalue is the same, and no further elaboration is made here.

[0103] In a specific implementation of the embodiment of the present invention, a normalized value of the product of the grayscale value and the grayscale gradient value of each pixel is used as the grayscale feature value of each pixel.

[0104] In a specific implementation of the embodiment of the present invention, the grayscale feature value acquisition process is expressed as follows: ;in, Pixel Grayscale eigenvalue of ; Pixel Gray value of Pixel Grayscale gradient; It is a normalization function, and all normalization functions in the embodiments of the present invention adopt linear normalization, which will not be further described later. It should be noted that in addition to obtaining the grayscale feature value by the normalized value of the product between the grayscale value and the grayscale gradient value, the implementer can also calculate it by other methods according to the specific implementation environment, such as taking the average between the normalized value of the grayscale value and the normalized value of the grayscale gradient value as the grayscale feature value.

[0105] Step S403: on each edge line segment, the degree of extrusion depression of each edge pixel point is determined according to the grayscale characteristic value distribution deviation between the first adjacent pixel point sequence and the second adjacent pixel point sequence corresponding to each edge pixel point.

[0106] The order of the sequence is the order of extension of edge pixels to both sides, and the real extruded edge is usually the boundary between the concave area and the normal area. According to the changes in the grayscale and grayscale gradient of the concave area, that is, the grayscale eigenvalue has a certain trend, and the grayscale and grayscale gradient of the normal area are relatively stable, that is, the grayscale eigenvalue is relatively stable. Therefore, the grayscale eigenvalue distribution deviation of the pixel point sequence on both sides of each edge pixel in the real extruded edge is usually large; therefore, by analyzing the grayscale eigenvalue distribution deviation of the pixel point sequence on both sides of each pixel point, the degree of extrusion concavity of each edge pixel point is further determined.

[0107] Preferably, in some possible implementations of the embodiments of the present invention, the specific implementation process of the method for obtaining the degree of extrusion depression includes:

[0108] Sequentially taking the first adjacent side pixel point sequence and the second adjacent side pixel point sequence as the target pixel point sequence;

[0109] According to the grayscale feature value difference between each pixel in the target pixel sequence and the previous pixel, the grayscale feature change value of each pixel in the target pixel sequence is determined; the grayscale feature change values ​​of all pixels in the target pixel sequence are averaged to obtain the overall change feature value of the target pixel sequence. Since the overall change feature value of the first adjacent pixel sequence is obtained in the same way as the overall change feature value of the second adjacent pixel sequence, the target pixel sequence is used to replace the first adjacent pixel sequence and the second adjacent pixel sequence for calculation; and the mean can characterize the overall size of a set of data, so the mean of the grayscale feature change values ​​of all pixels in the target pixel sequence is used as the corresponding overall change feature value. Further, according to the difference between the overall change feature value of the first adjacent pixel sequence and the overall change feature value of the second adjacent pixel sequence, the degree of extrusion and depression of each edge pixel on each edge segment is determined.

[0110] First, the change trend of the grayscale value and grayscale gradient in the concave area shows a decreasing trend from the outside to the inside, that is, in the direction from the outside to the inside, the corresponding grayscale feature change value is larger overall; while the grayscale value and grayscale gradient of the normal area without concave are relatively stable, and the corresponding grayscale feature change value is smaller overall. The real extruded edge is located at the boundary between the concave area and the normal area, so the overall change feature value of the pixel points on one side of the edge segment corresponding to the real extruded edge is larger, which is the side of the concave area; while the overall change feature value of the pixel points on the other side is smaller, which is the side of the normal area; therefore, for each edge pixel point, the greater the difference between the overall change feature value of the first adjacent side pixel point sequence and the overall change feature value of the second adjacent side pixel point sequence, the more it conforms to the characteristics of the pixel points on the real extruded edge, that is, the greater the corresponding extrusion concave degree, the more likely the edge segment where the edge pixel point is located is the real extruded edge.

[0111] In a specific implementation of the embodiment of the present invention, the process of obtaining the overall change characteristic value of the target pixel sequence is expressed by the formula: ;in, is the target pixel sequence The overall change characteristic value of is the target pixel sequence The number of pixels; is the target pixel sequence Middle Grayscale feature value of each pixel; is the target pixel sequence Middle Grayscale feature value of each pixel; is the absolute value symbol; is the target pixel sequence Middle The grayscale feature change value of each pixel.

[0112] Step S404: normalizing the average of the extrusion and depression degrees of all edge pixels on each edge segment to obtain the extrusion and depression coefficient of each edge segment.

[0113] The larger the mean of the extrusion depression degree of all edge pixels on the edge segment, the larger the deviation of the grayscale eigenvalues ​​on both sides of the edge segment. According to the characteristics of the grayscale distribution on both sides of the real extruded edge, the corresponding edge segment is more likely to be a real extruded edge, that is, the larger the corresponding extrusion depression coefficient.

[0114] In a specific implementation of the embodiment of the present invention, the process of obtaining the extrusion depression coefficient is expressed by the formula: ;in, For the The extrusion concavity coefficient of the edge segment; For the The number of edge pixels on an edge segment; For the The edge segment The overall change characteristic value of the first adjacent pixel sequence of the edge pixel; For the The edge segment The overall change characteristic value of the second adjacent side pixel sequence of the edge pixel point. For the The edge segment The degree of extrusion and depression of edge pixels; is the normalization function.

[0115] Step S104: determining the possibility of an extruded edge for each edge segment based on the edge deformation coefficient and the extrusion depression coefficient; screening out the real extruded edges in all edge segments according to the extruded edge possibility; and performing visual monitoring of the lamination production of fan-shaped generators according to the real extruded edges.

[0116] The larger the edge deformation coefficient and the larger the extrusion concave coefficient, the greater the possibility that the corresponding edge segment belongs to the real extruded edge. Therefore, the edge deformation coefficient and the extrusion concave coefficient are positively correlated with the possibility of the extruded edge.

[0117] Preferably, in some possible implementations of the embodiments of the present invention, the specific implementation process of the method for obtaining the possibility of extruding the edge in step S104 includes:

[0118] The product of the edge deformation coefficient and the extrusion concavity coefficient is normalized to determine the extrusion edge possibility of each edge segment.

[0119] In other possible implementations of the embodiments of the present invention, the possibility of edge extrusion may also be obtained in other ways, such as taking the average value between the normalized value of the edge deformation coefficient and the extrusion concave coefficient as the possibility of edge extrusion.

[0120] In a specific implementation of the embodiment of the present invention, the process of obtaining the possibility of extruding the edge is expressed by the formula: ; For the The possibility of extruding edges for edge segments; For the The extrusion concavity coefficient of the edge segment; For the The edge deformation coefficient of the edge segment.

[0121] Preferably, in some possible implementations of the embodiments of the present invention, the specific implementation process of the method for obtaining the real extruded edge in step S104 includes:

[0122] The edge line segment whose corresponding extrusion edge possibility is greater than the preset edge threshold is taken as the real extrusion edge. In a specific implementation of the embodiment of the present invention, the preset edge threshold is set to 0.8, and the implementer can adjust it according to the specific implementation environment.

[0123] Preferably, in some possible implementations of the embodiments of the present invention, the specific implementation process of visual monitoring of the lamination production of the fan-shaped generator according to the real extrusion edge in step S104 includes:

[0124] When the number of real extruded edges in the grayscale image of the fan-shaped piece surface is greater than or equal to the preset scrap threshold, the corresponding generator fan-shaped piece is determined to be a scrapped fan-shaped piece; when the number of real extruded edges in the grayscale image of the fan-shaped piece surface is less than the preset scrap threshold, the corresponding generator fan-shaped piece is determined to be a normal fan-shaped piece. In a specific implementation of an embodiment of the present invention, the preset scrap threshold is set to 3, and the implementer can adjust it according to the specific implementation environment.

[0125] In summary, the visual monitoring method for the stacking production of fan-shaped generators based on image processing determines the edge deformation coefficient corresponding to each edge segment according to the extension performance of the extruded edge under the influence of extrusion stress; determines the extrusion depression coefficient corresponding to each edge segment based on the grayscale performance of the extrusion depression; thereby comprehensively representing the extrusion edge possibility of each edge segment through the edge deformation coefficient and the extrusion depression coefficient, making the extrusion edge of the motor fan-shaped piece represented by the real extrusion edge screened out according to the extrusion edge possibility more accurate, thereby making the visual monitoring effect of the stacking production of fan-shaped generators better.

[0126] This application also provides a visual monitoring system for the stacking production of fan-shaped generators based on image processing. Figure 5 , which shows a structural diagram of a fan-shaped slice generator stacking production visual monitoring system based on image processing provided by an embodiment of the present invention, the system includes: a data acquisition preprocessing module 501, a first determination module 502, a second determination module 503 and a fan-shaped slice generator stacking production visual monitoring module 504.

[0127] The data acquisition preprocessing module 501 is used to perform edge detection based on the grayscale image of the surface of the generator fan-shaped pieces after the fan-shaped pieces are stacked, and determine the fan-shaped piece edges and all edge segments;

[0128] A first determination module 502 is used to determine the edge deformation coefficient of each edge line segment based on the influence of the extrusion stress of the sector piece and according to the extension direction of each edge line segment, the straight line similarity and the distance between each edge line segment and the edge of the sector piece;

[0129] The second determination module 503 is used to determine the extrusion depression coefficient of each edge line segment based on the grayscale representation of the extrusion depression on the surface grayscale image of the sector piece and the grayscale change trend and gradient change trend within the neighborhood of each edge line segment;

[0130] The fan-shaped blade generator stacking production visual monitoring module 504 is used to determine the extrusion edge possibility of each edge segment based on the edge deformation coefficient and the extrusion depression coefficient; screen out the real extrusion edges in all edge segments according to the extrusion edge possibility; and perform visual monitoring of the fan-shaped blade generator stacking production according to the real extrusion edges.

[0131] It should be noted that the system provided in the above embodiment is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device is divided into different functional modules to complete all or part of the functions described above. In addition, the visual monitoring system for the stacking production of fan-shaped generators based on image processing and the visual monitoring method embodiment for the stacking production of fan-shaped generators based on image processing provided in the above embodiment belong to the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0132] The present application also provides a computer device. Figure 6 , which shows a schematic diagram of the structure of a computer device provided by an embodiment of the present invention, the computer device includes a memory 601, a processor 602, and a computer program 603 stored in the memory 601 and running on the processor 602, wherein when the processor 602 executes the computer program 603, the computer device can execute any one of the above-mentioned visual monitoring methods for the stacking production of fan-shaped generators based on image processing.

[0133] An embodiment of the present application also provides a computer program product. When the computer program product is run on a computer device, the computer device can execute any one of the above-mentioned visual monitoring methods for the stacking production of fan-shaped generators based on image processing.

[0134] An embodiment of the present application also provides a computer-readable storage medium, in which a computer program code is stored. When the computer program code is executed on a computer device, the computer device can execute any one of the above-mentioned visual monitoring methods for the stacking production of fan-shaped generators based on image processing.

[0135] In the embodiments provided in the present application, it should be understood that the provided computer device, computer program product and computer-readable storage medium are all used to execute the corresponding methods provided above. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the methods provided above and will not be repeated here.

[0136] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0137] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A visual monitoring method for the lamination production of fan-shaped generators based on image processing, characterized in that: The method comprises: Perform edge detection based on the grayscale image of the surface of the generator fan-shaped pieces after the fan-shaped pieces are stacked, and determine the fan-shaped piece edges and all edge segments; Based on the influence of the extrusion stress of the fan-shaped sheet, the edge deformation coefficient of each edge line segment is determined according to the extension direction of each edge line segment, the straight line similarity and the distance between each edge line segment and the edge of the fan-shaped sheet; On the grayscale image of the surface of the sector sheet, based on the grayscale representation of the extrusion depression, and according to the grayscale change trend and the gradient change trend within the neighborhood of each edge segment, the extrusion depression coefficient of each edge segment is determined; Based on the edge deformation coefficient and the extrusion depression coefficient, the possibility of the extrusion edge of each edge line segment is determined; according to the possibility of the extrusion edge, the real extrusion edges in all edge line segments are screened out; according to the real extrusion edges, the visual monitoring of the lamination production of the fan-shaped generator is performed; The method for obtaining the edge deformation coefficient includes: Based on the shortest distance from each edge pixel point on each edge line segment to the edge of the fan-shaped piece, determine the edge distance of each edge pixel point; According to the number of edge pixels, the edge distance range and the minimum edge distance on each edge segment, the position direction squeezing index of each edge segment is determined; the edge distance range is positively correlated with the position direction squeezing index, and the number of edge pixels and the minimum edge distance are negatively correlated with the position direction squeezing index; Determine the edge smoothness corresponding to each edge line segment, which represents the smoothness of the edge contour, and the straight line similarity corresponding to each edge line segment, which represents the similarity of the fitted straight lines; Based on the position direction extrusion index, the edge smoothness and the straight line similarity, the edge deformation coefficient of each edge segment is determined; the position direction extrusion index, the edge smoothness and the straight line similarity are all positively correlated with the edge deformation coefficient.

2. The visual monitoring method for stacking production of fan-shaped generators based on image processing according to claim 1 is characterized in that: The method for obtaining the position direction squeezing index includes: Determine an edge pointing coefficient of each edge line segment according to a ratio of the edge distance extreme difference to the number of edge pixels; The position direction extrusion index of each edge line segment is determined by multiplying the negative correlation mapping value of the minimum edge distance value by the edge orientation coefficient.

3. The visual monitoring method for stacking production of fan-shaped generators based on image processing according to claim 1 is characterized in that: The method for obtaining the edge smoothness degree includes: Based on the chain code method, the chain code corresponding to each edge segment is counted; on the chain code corresponding to each edge segment, the difference between each chain code value and the next adjacent chain code value is used as the local chain code change value of each chain code value; The edge smoothness of each edge segment is determined according to the negative correlation mapping value of the mean of the local chain code change values ​​of all chain code values ​​on the chain code corresponding to each edge segment.

4. The visual monitoring method for lamination production of fan-shaped generators based on image processing according to claim 1 is characterized in that: The method for obtaining the straight line similarity degree comprises: Determine the fitting straight line corresponding to each edge segment based on a straight line fitting method; Perform dynamic time warping on each edge segment and the corresponding fitting line to determine the DTW distance between each edge segment and the corresponding fitting line; The DTW distance is subjected to negative correlation mapping processing to determine the straight line similarity of each edge line segment.

5. The visual monitoring method for stacking production of fan-shaped generators based on image processing according to claim 4 is characterized in that: The method for obtaining the extrusion depression coefficient includes: On each edge line segment, a preset number of neighboring pixel points that are closest to each edge pixel point in a preset first reference direction are arranged according to the preset first reference direction to determine a first adjacent pixel point sequence for each edge pixel point; on each edge line segment, a preset number of neighboring pixel points that are closest to each edge pixel point in a preset second reference direction are arranged according to the preset second reference direction to determine a second adjacent pixel point sequence for each edge pixel point; wherein both the preset first reference direction and the preset second reference direction are perpendicular to the fitting straight line of the edge line segment, and the preset first reference direction and the preset second reference direction are opposite; Determine the grayscale characteristic value of each pixel point according to the grayscale value and grayscale gradient value of each pixel point; the grayscale value and grayscale gradient value are both positively correlated with the grayscale characteristic value; On each edge line segment, the degree of extrusion depression of each edge pixel point is determined according to the grayscale characteristic value distribution deviation between the first adjacent side pixel point sequence and the second adjacent side pixel point sequence corresponding to each edge pixel point; The average of the extrusion and depression degrees of all edge pixels on each edge segment is normalized to obtain the extrusion and depression coefficient of each edge segment.

6. The visual monitoring method for stacking production of fan-shaped generators based on image processing according to claim 5 is characterized in that: The method for obtaining the degree of extrusion depression includes: sequentially taking the first adjacent side pixel point sequence and the second adjacent side pixel point sequence as a target pixel point sequence; Determine the grayscale feature change value of each pixel in the target pixel sequence according to the grayscale feature value difference between each pixel and the previous pixel in the target pixel sequence; The grayscale feature change values ​​of all pixels in the target pixel sequence are averaged to obtain the overall change feature value of the target pixel sequence; The degree of extrusion and depression of each edge pixel point on each edge line segment is determined according to the difference between the overall change characteristic value of the first adjacent side pixel point sequence and the overall change characteristic value of the second adjacent side pixel point sequence.

7. The visual monitoring method for lamination production of fan-shaped generators based on image processing according to claim 1 is characterized in that: The method for obtaining the possibility of extruding the edge includes: The product of the edge deformation coefficient and the extrusion concave coefficient is normalized to determine the extrusion edge possibility of each edge segment.

8. The visual monitoring method for lamination production of fan-shaped generators based on image processing according to claim 1 is characterized in that: The method for obtaining the real extruded edge includes: The edge line segment whose corresponding extrusion edge possibility is greater than the preset edge threshold is regarded as the real extrusion edge.

9. The visual monitoring method for stacking production of fan-shaped generators based on image processing according to claim 1 is characterized in that: The method for visually monitoring the lamination production of fan-shaped generators according to the real extrusion edge comprises: When the number of real extruded edges in the grayscale image of the surface of the fan-shaped piece is greater than or equal to a preset scrap threshold, the corresponding generator fan-shaped piece is determined to be a scrapped fan-shaped piece; When the number of real extruded edges in the grayscale image of the surface of the sector piece is less than a preset scrap threshold, the corresponding generator sector piece is determined to be a normal sector piece.

Images