Aluminum material precision cutting method and system based on machine vision

Abstract

The present invention relates to the field of image processing technology, and in particular to a method and system for precise cutting of aluminum materials based on machine vision. The method comprises the steps of: obtaining an image of an aluminum material, preprocessing it, and obtaining a plurality of superpixel blocks; obtaining the clustering result of the index position of all pixel points corresponding to any gray value in any superpixel block; calculating the significance of the gray value according to the Euclidean distance between all pixel points and the cluster center of the cluster where they are located and the information entropy of all gray values; calculating the degree of outlier of any pixel point; calculating the characteristic intensity of the pixel point according to the degree of outlier of any pixel point, the significance of the gray value, and the gray difference of the pixel point, and calculating the filter window side length of the pixel point according to the characteristic intensity; completing filtering according to the filter window side length, and completing precise cutting of the aluminum material according to the filtered image. The present invention is more accurate in cutting aluminum materials and avoids waste of resources.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to a method and system for accurately cutting aluminum materials based on machine vision. Background Art

[0002] Aluminum profiles are made of aluminum rods through hot melting and extrusion to obtain aluminum profiles of different cross-sectional shapes. Due to the characteristics of low density, high strength and good plasticity, aluminum profiles are widely used in the industrial field. Through cutting technology, aluminum profiles can be cut into various shapes and sizes to meet different needs.

[0003] If there are defects in the metal sheet, directly cutting it will directly affect the quality of the sheet. Therefore, it is necessary to identify the defects of the sheet so that when cutting it, the defective part can be avoided to improve the quality of the cut sheet and avoid wasting resources.

[0004] In aluminum materials, there are not only defective areas, but also the edges of aluminum materials need to be considered, so that the cutting path of the cutting knife can be reasonably planned according to the defective areas and edge areas, so as to maximize the use of aluminum materials and avoid waste of resources. Since defective areas and edge areas usually show different characteristics from the surrounding areas, and edge detection algorithms can capture the most obvious grayscale or color changes in the image, the edge detection algorithm can be used to obtain defective areas and edge areas of aluminum materials.

[0005] Since aluminum materials may produce some subtle features of uneven surfaces during the production process, these subtle features of uneven surfaces may be mistakenly identified as defect areas or edge areas during edge detection of aluminum materials. Therefore, it is necessary to filter the aluminum material image to blur the subtle features of uneven surfaces. Mean filtering adjusts the pixel value by calculating the average value of each pixel neighborhood in the image, which helps to reduce subtle features of uneven surfaces. However, since the traditional mean filtering algorithm is a preset fixed-size filter window, if the filter window is set too small, it will not be able to effectively eliminate subtle features of uneven surfaces. If the filter window is set too large, it will blur the defect areas and edge areas of the aluminum material, making the cutting results of the aluminum material inaccurate and causing a waste of resources. Summary of the invention

[0006] The purpose of the present invention is to provide a method and system for accurately cutting aluminum materials based on machine vision, so as to solve the technical problem of waste of resources caused by inaccurate cutting of aluminum materials.

[0007] In the first aspect, the present invention provides a method for accurately cutting aluminum materials based on machine vision, which adopts the following technical solutions:

[0008] Acquire an aluminum material image, pre-process the aluminum material image, and obtain a plurality of super-pixel blocks;

[0009] Obtain the index positions of all pixel points corresponding to any grayscale value in any superpixel block, cluster the index positions, and obtain clustering results; calculate the significance of the grayscale value according to the Euclidean distance between all pixel points and the cluster center of the cluster where they are located and the information entropy of all grayscale values; the significance is negatively correlated with the Euclidean distance and positively correlated with the information entropy;

[0010] For any pixel point, obtain the standard deviation of the Euclidean distances between all pixels of the cluster where the pixel point is located and the cluster center of the cluster, and calculate the outlier degree of the pixel point according to the standard deviation and the Euclidean distance between the pixel point and the cluster center, wherein the outlier degree is positively correlated with the Euclidean distance and negatively correlated with the standard deviation;

[0011] The sum of the absolute values ​​of the grayscale value differences between any pixel point and its eight neighboring pixels is taken as the local difference of the pixel point, the significance of the grayscale value of the pixel point and the ratio of the product of the local difference to the outlier degree are taken as the feature intensity of the pixel point, and the side length of the filter window of the pixel point is calculated according to the feature intensity, and the size of the filter window is negatively correlated with the feature intensity; filtering is completed according to the side length of the filter window, and precise cutting of aluminum materials is completed according to the filtered image.

[0012] The effect is that the present invention calculates the feature intensity of any pixel point, and adaptively obtains the filter side length of the pixel point according to the feature intensity of the pixel point, so that the filter side length of the pixel point with stronger feature intensity is smaller, and the features of the pixel point with stronger features can be retained to the maximum extent, thereby making the filtering result more accurate, and further making the precise cutting result of the aluminum material more accurate.

[0013] Preferably, the preprocessing of the aluminum material image to obtain a plurality of super pixel blocks comprises the steps of:

[0014] The aluminum material image is segmented using a superpixel segmentation algorithm to obtain multiple superpixel blocks.

[0015] Preferably, the clustering result includes:

[0016] The index positions of all pixel points are clustered using an iterative self-organizing clustering algorithm to obtain a clustering result, wherein the clustering result includes: the index positions of all pixel points in each cluster and the cluster center of each cluster.

[0017] Preferably, the significance satisfies the expression:

[0018] ;

[0019] In the formula, Indicates The first superpixel block The significance of the gray value; Indicates The information entropy of all gray values ​​of a superpixel block; Indicates The total number of clusters clustered according to the index position of the gray value; Indicates the cluster number; Indicates The total number of pixels in the clusters; Indicates the pixel ordinal number; Indicates The Euclidean distance between a pixel and the cluster center of the cluster to which it belongs; Expressed as a natural constant The exponential function of base .

[0020] The effect is that for all pixels of any gray value, by clustering the index positions of all pixels, and according to the Euclidean distance between all pixels and the cluster center of the cluster in which they are located in the clustering result, the concentration of the gray value in the superpixel block can be calculated. If the Euclidean distance is smaller, the concentration is higher, that is, all pixels corresponding to the gray value have strong spatial consistency in the superpixel block; information entropy reflects the complexity of all gray values ​​in the superpixel block. The higher its value, the more complex the gray distribution in the superpixel block, which helps to emphasize the important change areas in the image. Combining these two, the influence of noise on the calculation of gray value significance can be effectively reduced.

[0021] Preferably, the outlier degree satisfies the expression:

[0022] ;

[0023] In the formula, Indicates The gray value The first The degree of outlier of each pixel; Indicates The gray value The first cluster The Euclidean distance between a pixel and the cluster center of the cluster to which it belongs; Indicates The total number of pixels in the clusters; Indicates The average distance between all pixels in a cluster and the cluster center of the cluster; Represents the standard normalization function.

[0024] The effect is that by calculating the degree of outlier of any pixel in any grayscale value, the relationship between the pixel and its grayscale value can be quantified. If the degree of outlier is smaller, it means that the characteristic intensity of the pixel is closer to the significance of the grayscale value of the pixel, thereby making the calculation result of the characteristic intensity of subsequent pixels more accurate.

[0025] Preferably, the side length of the filtering window satisfies the expression:

[0026] ;

[0027] In the formula, Indicates the first The side length of the filter window of pixels; Indicates the side length of the preset initial filtering window; Indicates the first The characteristic intensity of each pixel; Expressed as a natural constant The exponential function with base ; Represents the ceiling function.

[0028] The effect is that the filter window side length of any pixel is calculated according to the feature intensity of the pixel, so that the filter window side length of the pixel with larger feature intensity is smaller, which can retain the feature of the pixel to a greater extent, thereby making the filtering result of the aluminum material image more accurate.

[0029] Preferably, the local difference satisfies the expression:

[0030] ;

[0031] In the formula, Indicates The gray value The first cluster The local difference of pixels; Represents the ordinal number of the neighboring pixels of any pixel; Indicates The gray value The first cluster The gray value of each pixel; Indicates The gray value The first cluster The first pixel in the neighborhood of The gray value of each pixel; Represents the absolute value symbol.

[0032] The effect is that local difference refers to the grayscale change between adjacent pixels. By calculating the local difference of pixel points, the key feature areas in the image can be identified more accurately, and the areas with significant grayscale changes in the image can be highlighted, so that the calculation results of the characteristic intensity of the pixel points are more obvious, and the filtering results of the aluminum material image are further made more accurate.

[0033] In a second aspect, the present invention provides an aluminum material precision cutting system based on machine vision, which adopts the following technical solutions:

[0034] The machine vision-based aluminum material precision cutting system comprises: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the machine vision-based aluminum material precision cutting method is implemented.

[0035] By adopting the above technical solution, the above-mentioned machine vision-based aluminum material precision cutting method is generated into a computer program and stored in a memory to be loaded and executed by a processor, so that a terminal device is manufactured based on the memory and the processor for easy use.

[0036] The present invention has the following technical effects:

[0037] The present invention obtains multiple super-pixel blocks by segmenting an image. For any grayscale value in the super-pixel block, the index positions of all corresponding pixel points are clustered to obtain a clustering result. The significance of the grayscale value is calculated based on the clustering result and the distribution of the grayscale value in the super-pixel block, thereby making the significance calculation result of the grayscale value more accurate.

[0038] Furthermore, the present invention comprehensively calculates the characteristic strength of the pixel point based on the significance of the grayscale value of the pixel point, the degree of outlier of the pixel point, and the local difference, and adaptively obtains the filter side length of the pixel point based on the characteristic strength of the pixel point, so that the filter side length of the pixel point with stronger characteristic strength is smaller, and the characteristics of the pixel point with stronger characteristics can be retained to the maximum extent, thereby making the filtering result more accurate, and further making the precise cutting result of the aluminum material more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] By reading the detailed description below with reference to the accompanying drawings, the above and other purposes, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-restrictive manner, and the same or corresponding reference numerals represent the same or corresponding parts.

[0040] Figure 1 It is a flow chart of a method for precise cutting of aluminum materials based on machine vision in an embodiment of the present invention. DETAILED DESCRIPTION

[0041] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0042] It should be understood that when the terms "first", "second", etc. are used in the claims, descriptions, and drawings of the present invention, they are only used to distinguish different objects, rather than to describe a specific order. The terms "include" and "comprise" used in the description and claims of the present invention indicate the presence of the described features, wholes, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components, and / or their collections.

[0043] The embodiment of the present invention discloses a method for accurately cutting aluminum materials based on machine vision, referring to Figure 1 , comprising steps S1 to S5:

[0044] S1: Acquire an aluminum material image, pre-process the aluminum material image, and obtain multiple super-pixel blocks;

[0045] The image of the aluminum material is photographed by a camera to obtain an RGB image of the aluminum material, which is gray-scaled to obtain a grayscale image of the aluminum material. The grayscale image of the aluminum material is segmented using a superpixel segmentation algorithm to obtain multiple superpixel blocks.

[0046] It should be noted that the superpixel segmentation algorithm is an image segmentation algorithm that is used to divide an image into many coherent, smaller regions, which are called superpixels. The principle is to classify pixels in the image into superpixels with similar features through clustering methods to improve segmentation efficiency and quality. The algorithm first initializes the pixel groups based on the image's color, texture, and position features, and then builds a graphical model of the image based on these features. For example, the commonly used SLIC (Simple Linear Iterative Cluster) algorithm evenly distributes the initial superpixel centers in the image and continuously adjusts the positions of these centers in an iterative manner to minimize the color and texture differences within the superpixel. The pixels within each superpixel area will be pulled closer to its center point while maintaining the coherence and shape uniformity of the area.

[0047] At this point, the preprocessing of the aluminum material image has been completed, and multiple superpixel blocks have been obtained.

[0048] S2: For any superpixel block, calculate the significance of any gray value in the superpixel block;

[0049] It should be noted that when cutting aluminum materials, a superpixel segmentation algorithm is used to obtain multiple superpixel blocks, each of which is a local area of ​​the image. The grayscale distribution in each superpixel block is relatively uniform. Since the boundaries and details in the aluminum material image are key information, calculating the significance of any grayscale value in each superpixel block helps to accurately locate the boundaries, especially in places with obvious grayscale changes.

[0050] It should be further explained that, for any grayscale value, the distribution of all pixel points corresponding to the grayscale value in the superpixel will also affect the significance of the grayscale value. If the distribution of the grayscale value in the superpixel block is more concentrated, the significance of the grayscale value will be higher. Therefore, it is necessary to analyze the distribution of the grayscale value in the superpixel block.

[0051] In one embodiment, for any grayscale value in any superpixel block, the index positions of all pixel points corresponding to the grayscale value are obtained, and the index positions of all pixel points corresponding to the grayscale value are clustered using an iterative self-organizing clustering algorithm to obtain multiple clusters and the cluster center of each cluster.

[0052] Calculate the significance of any gray value based on the clustering results:

[0053] ;

[0054] In the formula, Indicates The first superpixel block The significance of the gray value; Indicates The information entropy of all gray values ​​of a superpixel block; Indicates The total number of clusters clustered according to the index position of the gray value; Indicates the cluster number; Indicates The total number of pixels in the clusters; Indicates the pixel ordinal number; Indicates The Euclidean distance between a pixel and the cluster center of the cluster to which it belongs; Expressed as a natural constant The exponential function of base .

[0055] in, Indicates The information entropy of all gray values ​​of a pixel block. The information entropy can reflect the complexity of the pixel block. The larger its value is, the more irregular the gray value in the super pixel block is, and the easier it is to be noticed, that is, the greater the significance of the gray value in the super pixel block is. Indicates The first superpixel block The smaller the value is, the smaller the discrete degree of the gray value is, that is, the more concentrated the gray value is, the stronger the spatial consistency of the gray value in the super pixel block is, that is, the higher the significance of the gray value is.

[0056] At this point, the significance calculation of any gray value is completed.

[0057] S3: Calculate the outlier degree of any pixel of any gray value;

[0058] It should be noted that when enhancing the image of aluminum materials, it is necessary to analyze each pixel in the aluminum materials so as to adaptively enhance the image according to the characteristic strength of each pixel. For pixels with stronger characteristic strength, a smaller filter window is used to retain the characteristics of the pixel. For any pixel, the significance of its gray value can reflect the characteristic strength of the pixel. The higher the significance of the gray value of the pixel, the higher the characteristic strength of the pixel. The farther the pixel is from all the pixels corresponding to the gray value, that is, the higher the degree of outlier of the pixel, the greater the difference between the characteristic strength of the pixel and the significance of the gray value. Therefore, it is necessary to calculate the degree of outlier of the pixel in order to calculate the characteristic strength of the pixel later.

[0059] In one embodiment, the outlier degree calculation formula of any pixel point of any gray value is as follows:

[0060] ;

[0061] In the formula, Indicates The clustering results of gray values The first cluster The degree of outlier of each pixel; Indicates The clustering results of gray values The first cluster The Euclidean distance between a pixel and the cluster center of the cluster to which it belongs; Indicates The total number of pixels in the clusters; Indicates The average distance between all pixels in a cluster and the cluster center of the cluster; Represents the standard normalization function.

[0062] In the formula, Represents the Euclidean distance between the pixel and its cluster. The larger the value, the higher the degree of outlier of the pixel. It represents the standard deviation of the Euclidean distance between all pixels of the cluster where the pixel is located and the cluster center of the cluster. The smaller the value, the more concentrated all pixels of the cluster where the pixel is located are, and the higher the confidence level of the discrete degree of the pixel is.

[0063] At this point, the outlier degree calculation of any pixel point of any gray value is completed.

[0064] S4: Calculate the feature intensity of any pixel point, and calculate the side length of the filter window of the pixel point according to the feature intensity of the pixel point;

[0065] In one embodiment, for any pixel, the sum of the absolute values ​​of the grayscale differences between the pixel and the pixels in its eight neighborhoods is taken as the local difference of the pixel, and the calculation formula is as follows:

[0066] ;

[0067] In the formula, Indicates The gray value The first cluster The local difference of pixels; Represents the ordinal number of the neighboring pixels of any pixel; Indicates The gray value The first cluster The gray value of each pixel; Indicates The gray value The first cluster The first pixel in the neighborhood of The gray value of each pixel; Represents the absolute value symbol.

[0068] The greater the grayscale difference between the pixel and all the pixels in its neighborhood, the greater the local difference of the pixel.

[0069] The significance of the gray value of the pixel, the product of the local difference and the ratio of the outlier degree are taken as the feature intensity of the pixel.

[0070] It should be noted that when using the mean filtering algorithm to filter all pixels in the aluminum material image, different filter window sizes will affect the image details. A larger filter window can effectively smooth the image and reduce the impact of random noise, while a smaller filter window can effectively retain the characteristics of the pixels in the image. Therefore, it is necessary to calculate the filter window side length of the pixel based on the characteristic intensity of the pixel.

[0071] The side length of the filter window of the pixel is calculated according to the characteristic intensity of the pixel. The calculation formula is as follows:

[0072] ;

[0073] In the formula, Indicates the first The side length of the filter window of pixels; Indicates the side length of the preset initial filtering window; Indicates the first The characteristic intensity of each pixel; Expressed as a natural constant The exponential function with base ; Represents the ceiling function. The length of the initial filtering window is not specifically limited in the present invention, and the implementer can set it according to the actual implementation situation, for example .

[0074] Among them, the stronger the feature intensity of the pixel point, the more it is necessary to retain the feature of the pixel point, and the smaller the filter window, the more the feature of the pixel point can be retained. Therefore, the stronger the feature intensity of the pixel point, the smaller the filter window of the pixel point.

[0075] At this point, the calculation of the filter window side length of any pixel point is completed.

[0076] S5: filtering the aluminum material image according to the side length of the filtering window, and accurately cutting the aluminum material according to the filtered image;

[0077] In one embodiment, for any pixel point, the pixel point is filtered using a mean filtering algorithm according to the calculated size of the filtering window of the pixel point to obtain a filtered aluminum material image. The filtered aluminum material image is subjected to edge detection using a Canny operator to obtain an edge detection result, wherein the edge detection result includes: a material edge area of ​​the aluminum material and a defect area of ​​the aluminum material.

[0078] The staff plans the cutting position of the cutting knife according to the edge area and the defect area, which further facilitates the staff to plan the cutting path of the cutting knife according to the edge area and the defect area in time, so as to improve the accuracy of cutting aluminum materials, improve the quality of the cut aluminum materials, and avoid waste of resources.

[0079] Through the above steps, the filtering of the aluminum material image is completed, and the precise cutting of the aluminum material is completed according to the filtered aluminum material image.

[0080] An embodiment of the present invention also discloses a precise cutting system for aluminum materials based on machine vision, including a processor and a memory, wherein the memory stores computer program instructions. When the computer program instructions are executed by the processor, a precise cutting method for aluminum materials based on machine vision according to the present invention is implemented.

[0081] The above system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface, and their configuration and functions are known in the art, so they will not be described in detail here.

[0082] In the present invention, the aforementioned memory may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium may be any suitable magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory, a dynamic random access memory, a static random access memory, an enhanced dynamic random access memory, a high bandwidth memory, a hybrid storage cube, etc., or any other medium that can be used to store the required information and can be accessed by an application, a module, or both. Any such computer storage medium may be part of the device or accessible or connectable to the device.

[0083] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

[0084] The above are all preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for precise cutting of aluminum materials based on machine vision, characterized in that: Includes steps: Acquire an aluminum material image, pre-process the aluminum material image, and obtain a plurality of super-pixel blocks; Get the index positions of all pixels corresponding to any gray value in any superpixel block, cluster the index positions, and obtain clustering results; calculate the significance of the gray value according to the Euclidean distance between all pixels and the cluster center of the cluster where they are located and the information entropy of all gray values, satisfying the expression: ; In the formula, Indicates The first superpixel block The significance of the gray value; Indicates The information entropy of all gray values ​​of a superpixel block; Expressing the The total number of clusters clustered according to the index position of the gray value; Indicates the cluster number; Indicates The total number of pixels in the clusters; Indicates the pixel ordinal number; Indicates The Euclidean distance between a pixel and the cluster center of the cluster to which it belongs; Expressed as a natural constant is an exponential function with base number; significance is negatively correlated with Euclidean distance and positively correlated with information entropy; For any pixel point, obtain the standard deviation of the Euclidean distances between all pixels of the cluster where the pixel point is located and the cluster center of the cluster, and calculate the outlier degree of the pixel point according to the standard deviation and the Euclidean distance between the pixel point and the cluster center, wherein the outlier degree is positively correlated with the Euclidean distance and negatively correlated with the standard deviation; The sum of the absolute values ​​of the grayscale value differences between any pixel and its eight neighboring pixels is taken as the local difference of the pixel, the significance of the grayscale value of the pixel and the ratio of the product of the local difference to the degree of outlier are taken as the feature strength of the pixel, and the side length of the filter window of the pixel is calculated according to the feature strength, and the side length of the filter window is negatively correlated with the feature strength; The filtering is completed according to the side length of the filtering window, and the precise cutting of the aluminum material is completed according to the filtered image.

2. The method for precise cutting of aluminum materials based on machine vision according to claim 1, characterized in that: The method of preprocessing the aluminum material image to obtain a plurality of super pixel blocks comprises the following steps: The aluminum material image is segmented using a superpixel segmentation algorithm to obtain multiple superpixel blocks.

3. The method for precise cutting of aluminum materials based on machine vision according to claim 1, characterized in that: The clustering results include: The index positions of all pixel points are clustered using an iterative self-organizing clustering algorithm to obtain a clustering result, wherein the clustering result includes: the index positions of all pixel points in each cluster and the cluster center of each cluster.

4. The method for precise cutting of aluminum materials based on machine vision according to claim 1, characterized in that: The outlier degree satisfies the expression: ； In the formula, Indicates The gray value The first The degree of outlier of each pixel; Indicates The gray value The first The Euclidean distance between a pixel and the cluster center of the cluster to which it belongs; Indicates The total number of pixels in the clusters; Indicates The average distance between all pixels in a cluster and the cluster center of the cluster; Represents the standard normalization function.

5. The method for precise cutting of aluminum materials based on machine vision according to claim 1, characterized in that: The filter window side length satisfies the expression: ； In the formula, Indicates the first The side length of the filter window of pixels; Indicates the side length of the preset initial filtering window; Indicates the first The characteristic intensity of each pixel; Expressed as a natural constant The exponential function with base ; Represents the ceiling function.

6. The method for precise cutting of aluminum materials based on machine vision according to claim 1, characterized in that: The local difference satisfies the expression: ； In the formula, Indicates The gray value The first The local difference of pixels; Represents the ordinal number of the neighboring pixels of any pixel; Indicates The gray value The first The gray value of each pixel; Indicates The gray value The first The first pixel in the neighborhood of The gray value of each pixel; Represents the absolute value symbol.

7. Aluminum material precision cutting system based on machine vision, characterized by: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a method for precise cutting of aluminum materials based on machine vision according to any one of claims 1 to 6 is implemented.

Images