Metal powder packaging quality detection system based on machine vision

Abstract

The invention discloses a metal powder packaging quality detection system based on machine vision, and relates to the technical field of visual inspection, the system comprises a visual acquisition module, an image processing module and a central processing module, the visual acquisition module is used for acquiring at least one group of detection images of each metal powder packaging bag; the visual acquisition module transmits the detection image module to the image processing module; after the image processing module receives the detection image, the detection image is matched with the corresponding technical powder packaging bag, and the image processing module preprocesses the detection image; and after receiving the preprocessed detection image, the central processing module performs visual identification on the detection image to obtain use quality data of the packaging bag. Through combination of the visual sensor module and the image processing module, accurate detection of the quality of the packaging bag is realized, and the packaging bag can be monitored in real time.

Description

Technical Field

[0001] The present invention relates to the technical field of visual inspection, and specifically to a metal powder packaging quality inspection system based on machine vision. Background Art

[0002] In the field of metal powder packaging, the quality control of packaging bags is a crucial step in ensuring product safety and performance. Metal powders are commonly used in industrial manufacturing fields such as 3D printing, metallurgy, etc., so their packaging must be able to maintain the purity, airtightness, and stability of the powder. In metal powder packaging, the packaging bags need to have high airtightness and compressive resistance to prevent external factors such as air and moisture from affecting the powder quality. To meet these requirements, the detection of quality parameters such as the airtightness and surface flatness of metal powder packaging bags has become crucial.

[0003] After retrieval, Chinese Patent (CN116309337B) discloses a packaging box quality inspection system based on image recognition. This patent performs segmentation processing on the top and side images of the obtained packaging box, and analyzes the segmented parts in sequence according to the size and difference of the defective area. During the analysis process, if a defective area appears or the difference in the defective areas that appear is large, subsequent analysis will no longer be carried out.

[0004] With the development of automation and intelligent technologies, more and more production lines have begun to integrate more advanced image recognition and analysis technologies. However, the existing technologies still face challenges such as overly single detection parameters and difficulty in providing real-time feedback on changes in packaging quality. They cannot provide real-time monitoring of the expansion, deformation, or position offset of packaging bags, and cannot effectively handle quality problems caused by changes in the position of the expanded part. Therefore, the present invention proposes a metal powder packaging quality inspection system based on machine vision. Summary of the Invention

[0005] The purpose of the present invention is to provide a metal powder packaging quality inspection system based on machine vision to solve the problems mentioned in the above background art.

[0006] The present invention can be achieved through the following technical solutions: A metal powder packaging quality inspection system based on machine vision, including a visual acquisition module, an image processing module, and a central processing module;

[0007] The visual acquisition module is used to acquire at least one set of detection images of each metal powder packaging bag, and the visual acquisition module transmits the detection image module to the image processing module;

[0008] After receiving the detected image, the image processing module matches the detected image with the corresponding technical powder packaging bag. When the visual acquisition module acquires multiple detected images, the image processing module establishes an image library for the corresponding packaging bag. The image library is used to store each detected image of the corresponding packaging bag, and the database distributes each detected image in a time series. Moreover, the image processing module preprocesses the detected image to improve its quality. The preprocessing includes denoising, enhancement, cropping, and scaling of the detected image.

[0009] After receiving the preprocessed detected image, the central processing module performs visual recognition on the detected image to obtain the usage quality data of the packaging bag. The usage quality data includes the continuity value C of the packaging bag edge e , the clarity C of the image elements q , the spatial correlation value D between the image elements p , and the unfolding degree T of the packaging bag surface s ;

[0010] The central processing module removes the unit of the usage quality data of the packaging bag and takes its numerical value, and calculates the quality score of the corresponding packaging bag through a formula. The formula used for the quality score is:

[0011]

[0012] In the formula, D r is the standard spatial correlation value of the image elements in the corresponding packaging bag, indicating the maximum tolerance deviation of the positions of the image elements when the packaging bag is in an ideal state;

[0013] T r is the standard surface unfolding degree of the corresponding packaging bag, representing the standard value of the surface unfolding of the packaging bag;

[0014] e is a constant to avoid the denominator being zero;

[0015] w1, w2, w3, and w4 are the weight coefficients of the continuity value C of the packaging bag edge e , the clarity C of the image elements q , the spatial correlation value D between the image elements p , and the unfolding degree T of the packaging bag surface s respectively;

[0016] The central processing module compares the quality score Q of the corresponding packaging bag with a preset standard score. When the quality score Q ≥ the standard score, the central processing module marks the corresponding packaging bag as qualified to facilitate normal operations on the marked packaging bag in subsequent processes;

[0017] If the quality score Q < the standard score, the central processing module marks the packaging bag as a suspicious package and conducts a secondary inspection manually later.

[0018] A further technical improvement of the present invention lies in that: the central processing module uses an edge detection algorithm for the detected image to identify the edges of the packaging bag in the detected image. The central processing module identifies the regions with large gray-scale changes in the detected image through the edge detection algorithm, and obtains the outer contour and the notch part of the packaging bag in the detected image through a contour extraction algorithm. And the central processing module calculates the continuity value C of the packaging bag edge through a formula e , and the formula used is:

[0019]

[0020] In the formula, L c is the length of the continuous part of the packaging bag; L t is the total edge length of the packaging bag; α is the weight coefficient of the number of notches, N g is the number of edge notches; β is the weight coefficient of the edge curvature of the packaging bag; K is the edge curvature of the packaging bag.

[0021] A further technical improvement of the present invention lies in that: the method for obtaining the clarity C of the image elements includes the following steps; q The visual acquisition module acquires the detected image, and the image processing module preprocesses the acquired detected image;

[0022] The central processing module calculates the brightness difference between each pixel point (x

[0023] , y i , y i ) in the preprocessed detected image and the previous pixel point (x i-1 , y i-1 );

[0024] And the central processing module accumulates the brightness difference values of each pair of adjacent pixel points to obtain the total brightness difference sum, and the central processing module calculates the clarity C of the image elements in the packaging bag by dividing the total brightness difference sum by the number of pixels n q , and the clarity C of the image elements q The calculation formula is:

[0025]

[0026] In the formula, I(x i , y i ) is the pixel value of the image at the position (x i , y i ); I(x i-1 , y i-1 ) is the pixel value of the image at the position (x i-1 , y i-1The pixel value of the previous pixel; n is the number of pixels in the calculation area.

[0027] A further technical improvement of the present invention lies in that: the image elements include a text part (such as production date, batch number, specification), a bar part (such as barcode, QR code), and a label part;

[0028] The central processing module calculates the clarity C of the image elements corresponding to the packaging bag q Before that, the content of each image element is recognized, and it is judged whether the content of the image element corresponding to the packaging bag image matches the preset content of the packaging bag image element, so as to determine whether the content of the image element corresponding to the packaging bag meets the predetermined standard or sequence;

[0029] If the matching is successful, the central processing module calculates the clarity C of the image element q Calculation.

[0030] A further technical improvement of the present invention lies in: the method for obtaining the spatial correlation value D between image elements, including: p The central processing module extracts the positions of each image element from the detected image through image processing, and each image element has a horizontal and vertical coordinate of the actual position (x

[0031] , y i , y i );

[0032] For each image element, the central processing module defines its target position in the packaging bag of the standard package according to the design or standard requirements of the corresponding packaging bag Target position Is a preset ideal position, representing the coordinates where these image elements should be in the packaging bag image;

[0033] For each image element, the central processing module calculates the difference between its actual position (x i , y i ) and the target position To obtain the spatial deviation of the image element, representing the difference between the actual position and the standard position;

[0034] Subsequently, the central processing module accumulates the spatial deviations of all image elements to obtain the overall spatial deviation, that is, the spatial correlation value D p , and the formula used is:

[0035]

[0036] In the formula, x i , y i Is the actual position coordinate of the i-th image element; is the target position coordinate of the i-th image element; m is the number of image elements.

[0037] A further technical improvement of the present invention lies in: the degree of unfolding T of the packaging bag surface s The acquisition method includes the following steps:

[0038] The central processing module calculates the preprocessed detection image to be converted into a grayscale image, and the central processing module averages the grayscale values h(x, y) of all pixel points in the grayscale image to calculate the average pixel value Mean(h) of the overall image. The formula for the average pixel value Mean(h) is:

[0039]

[0040] In the formula, W and H are the width and height of the image respectively;

[0041] Subsequently, the central processing module calculates the difference |h(x, y) - Mean(h)| between the grayscale value of each pixel point in the detected grayscale image and the average grayscale value of the overall image;

[0042] And the central processing module accumulates the grayscale differences of all pixel points and then divides by the total number of pixels W×H of the grayscale image to obtain the final surface unfolding degree T of the corresponding detection image s , and the specific formula is:

[0043]

[0044] A further technical improvement of the present invention lies in: the metal powder packaging quality detection system includes a pressing module, and the pressing module includes an execution unit and a pressure monitoring unit;

[0045] When the central processing module calculates the quality score Q of the corresponding packaging bag and the quality score Q ≥ the standard score, the execution unit is activated to apply a preset pressure to a preset part of the packaging bag, and the central processing module monitors the pressure applied by the execution unit to the packaging bag through the pressure monitoring unit;

[0046] Subsequently, the visual acquisition module acquires the recognition image of the corresponding packaging bag. The central processing module compares the recognition image with the detection image to obtain the deformation area and deformation amplitude of the packaging bag in the recognition image, and the central processing module detects the deformed part of the packaging bag through the formula. The formula is:

[0047]

[0048] In the formula, T d is the detection score of the deformed part, which is used to measure whether the position, quantity and amplitude of the expanded part of the packaging bag meet the expectations;

[0049] d is the number of deformed parts; A i is the area of the i-th deformed part, and the area of each deformed part is calculated through region analysis; A max is the maximum allowable deformation area, which is set based on the design standard and packing standard of the corresponding packaging bag and is used to measure whether each expanded part meets the allowable range; V i is the deformation amplitude of the i-th deformed part, which is measured by calculating the relative change amount of the deformed area between the detected image and the recognized image; V max is the maximum allowable deformation amplitude, which is the upper limit of the deformation amplitude specified during the design of the packaging bag and is used to ensure that the shape of the packaging bag meets the expectation; x i and y i are the coordinates of the i-th deformed part; is the target coordinate of the i-th deformed part, indicating the standard position where the deformed part should be located, usually set according to the packaging bag design or the target area; γ is the weight coefficient for restricting the coordinates of the deformed part and is used to adjust the influence degree of the offset of the expanded part from the standard position on the detection score of the deformed part;

[0050] And the central processing module compares the calculated detection score of the deformed part with the preset deformation score threshold. When the detection score of the deformed part < the deformation score threshold, the central processing module marks the corresponding packaging bag as qualified.

[0051] A further technical improvement of the present invention is that when the central processing module calculates the recognized image, the central processing module records the position and amplitude of the deformed part of the packaging bag and marks them as the initial position and the initial amplitude respectively;

[0052] Subsequently, the execution unit continuously applies pressure to the packaging bag, the visual acquisition module continuously acquires the recognized image of the corresponding packaging bag, and the central processing module arranges the continuously acquired recognized images in time series to obtain a set of recognized images;

[0053] The central processing module compares the position and amplitude of the deformed part of the packaging bag in the set of recognized images with the corresponding initial position and initial amplitude in time series, and respectively obtains the deformation position difference and the deformation amplitude difference;

[0054] Finally, the central processing module respectively compares the deformation position difference and the deformation amplitude difference with the preset position difference threshold and amplitude difference threshold;

[0055] When the deformation position difference < the position difference threshold and the deformation amplitude difference < the amplitude difference threshold, the central processing module marks the corresponding packaging bag as qualified.

[0056] Compared with the prior art, the present invention has the following beneficial effects:

[0057] Through the combination of the visual sensor module and the image processing module, the present invention realizes the precise detection of the quality of packaging bags, especially having significant advantages in detecting the swelling area, sealing performance and deformation. Compared with traditional methods, the present invention can conduct real-time monitoring of packaging bags, not only effectively capture the position and swelling amplitude of the swelling part of the packaging bag, but also monitor the dynamic changes of the swelling part, and adjust the production parameters through a feedback mechanism to ensure the consistency and stability of the quality of packaging bags;

[0058] Moreover, the present invention can detect quality elements such as the swelling part, sealing performance, and surface flatness of the packaging bag in real time, and generate a quality score through comprehensive evaluation, which can provide a comprehensive and accurate quality assessment, ensure that the quality of the packaging bag meets the standards, and avoid the problems of low detection accuracy and slow response in traditional methods;

[0059] On the other hand, through the positioning of the deformed part and the evaluation of the deformation amplitude, the present invention can ensure the sealing performance of the packaging bag, not only improve the detection accuracy, but also enhance the real-time and precision of the quality control of the packaging bag. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] For the convenience of those skilled in the art to understand, the present invention will be further described below in conjunction with the accompanying drawings.

[0061] Figure 1 is a flowchart of the present invention.

[0062] Figure 2 is a system block diagram of Embodiment 2 of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0063] To further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following will, in conjunction with the accompanying drawings and preferred embodiments, describe in detail the specific embodiments, structures, features and their effects according to the present invention.

[0064] Embodiment 1

[0065] Please refer to Figure 1 As shown, the present invention provides a metal powder packaging quality detection system based on machine vision, including a visual acquisition module, an image processing module, and a central processing module;

[0066] The visual acquisition module is used to acquire a set of detection images of each metal powder packaging bag, and the visual acquisition module transmits the detection image module to the image processing module;

[0067] After receiving the detection image, the image processing module matches the detection image with the corresponding technical powder packaging bag, and the image processing module preprocesses the detection image to improve the quality of the detection image. The preprocessing includes denoising, enhancement, cropping and scaling of the detection image;

[0068] After receiving the preprocessed detection image, the central processing module performs visual recognition on the detection image to obtain the usage quality data of the packaging bag. The usage quality data includes the continuous value C of the packaging bag edge e , the clarity C of the image elements q , the spatial correlation value D between the image elements p and the unfolding degree T of the packaging bag surface s ;

[0069] Among them, the image elements include the text part (such as production date, batch number, specification), the bar part (such as barcode, QR code) and the label part;

[0070] And before the central processing module calculates the clarity C of the image elements in the corresponding packaging bag q , it identifies the content of each image element and determines whether the content of the image element of the corresponding packaging bag matches the preset content of the packaging bag image element to determine whether the content of the image element of the corresponding packaging bag meets the predetermined standard or sequence;

[0071] If the matching is successful, the central processing module calculates the clarity C of the image elements q ;

[0072] If the matching fails, the packaging bag is marked as a suspicious package;

[0073] The central processing module uses an edge detection algorithm for the detection image to identify the edge of the packaging bag in the detection image. The central processing module identifies the area with large gray-scale changes in the detection image through the edge detection algorithm, and obtains the outer contour and notch part of the packaging bag in the detection image through the contour extraction algorithm. And the central processing module calculates the continuous value C of the packaging bag edge through a formula e , and the formula used is:

[0074]

[0075] In the formula, L c is the length of the continuous part of the packaging bag; L t is the total edge length of the packaging bag; α is the weight coefficient of the notch number, N g is the number of edge notches; β is the weight coefficient of the packaging bag edge curvature; K is the edge curvature of the packaging bag;

[0076] The higher the value of the continuous value C e , the more continuous the edge of the packaging bag is, and the better the packaging quality; the lower the score, the more defects there are on the edge of the packaging bag, and the worse the packaging quality;

[0077] The clarity C of the image elements qThe acquisition method includes the following steps;

[0078] The visual acquisition module acquires the detection image, and the image processing module preprocesses the acquired detection image;

[0079] For each pixel point (x i , y i ) in the preprocessed detection image, the central processing module calculates the brightness difference between it and the previous pixel point (x i-1 , y i-1 ). The brightness difference reflects the changes in edges or details in the image. A larger difference indicates richer details and higher clarity in the image;

[0080] And the central processing module accumulates the brightness difference values of each pair of adjacent pixel points to obtain the total brightness difference sum. And the central processing module calculates the clarity C of the image elements in the packaging bag by dividing the total brightness difference sum by the number of pixels n q , and the clarity C of the image elements q The calculation formula is:

[0081]

[0082] In the formula, I(x i , y i ) is the pixel value of the image at the position (x i , y i ); I(x i-1 , y i-1 ) is the previous pixel value of the image at the position (x i-1 , y i-1 ); n is the number of pixels in the calculation area;

[0083] The acquisition method of the spatial correlation value D p between image elements includes:

[0084] The central processing module extracts the positions of each image element from the detection image through image processing. Each image element has the horizontal and vertical coordinates (x i , y i ) of an actual position;

[0085] For each image element, the central processing module defines its target position in the packaging bag of the standard package according to the design or standard requirements of the corresponding packaging bag The target position is a preset ideal position, representing the coordinates where these image elements should be in the packaging bag image;

[0086] For each image element, the central processing module calculates its actual position (x i , yi ) The difference from the target position is used to obtain the spatial deviation of the image element, representing the difference between the actual position and the standard position;

[0087] Subsequently, the central processing module accumulates the spatial deviations of all image elements to obtain the overall spatial deviation, that is, the spatial correlation value D p , and the formula used is:

[0088]

[0089] In the formula, x i , y i are the actual position coordinates of the i-th image element; are the target position coordinates of the i-th image element; m is the number of image elements;

[0090] The degree of unfolding T of the packaging bag surface s The acquisition method includes the following steps:

[0091] The central processing module calculates the preprocessed detection image converted into a grayscale image, and the central processing module averages the grayscale values h(x, y) of all pixel points in the grayscale image to calculate the average pixel value Mean(h) of the overall image. The formula for the average pixel value Mean(h) is:

[0092]

[0093] In the formula, W and H are the width and height of the image respectively;

[0094] Subsequently, the central processing module calculates the difference |h(x, y) - Mean(h)| between the grayscale value of each pixel point in the detected grayscale image and the average grayscale value of the overall image;

[0095] And the central processing module accumulates the grayscale differences of all pixel points and then divides by the total number of pixels W×H of the grayscale image to obtain the final surface unfolding degree T of the corresponding detection image s , and the specific formula is:

[0096]

[0097] The smaller the degree of unfolding, the flatter the surface of the packaging bag. The larger the degree of unfolding, the more wrinkles or irregular shapes on the surface, which affect the quality of the packaging bag;

[0098] The central processing module removes the unit and takes the numerical value of the service quality data of the packaging bag, and calculates the quality score of the corresponding packaging bag through a formula. The formula for the quality score is:

[0099]

[0100] Wherein, D r is the standard spatial correlation value corresponding to the image elements in the packaging bag, representing the maximum tolerance deviation of the positions of the image elements when the packaging bag is in an ideal state;

[0101] T r is the standard surface unfolding degree corresponding to the packaging bag, representing the standard value of the surface unfolding of the packaging bag;

[0102] e is a constant to avoid a zero denominator;

[0103] w1, w2, w3, and w4 are the continuous values C at the edge of the packaging bag e , the clarity C of the image elements q , the spatial correlation value D between the image elements p and the unfolding degree T of the surface of the packaging bag s are the weight coefficients, which are obtained through historical data and experiments and are adjusted according to different packaging bags, the number of metal powder packages, and the transfer equipment to optimize the continuous value C at the edge of the packaging bag e , the clarity C of the image elements q , the spatial correlation value D between the image elements p and the unfolding degree T of the surface of the packaging bag s on the quality score;

[0104] The central processing module compares the quality score Q of the corresponding packaging bag with the preset standard score. When the quality score Q ≥ the standard score, the central processing module marks the corresponding packaging bag as qualified to facilitate the normal operation of the marked packaging bag in the subsequent process;

[0105] If the quality score Q < the standard score, the central processing module marks the packaging bag as a suspicious package, and subsequent secondary detection is performed manually.

[0106] Embodiment 2

[0107] A metal powder packaging quality detection system based on machine vision includes a vision acquisition module, an image processing module, and a central processing module;

[0108] The vision acquisition module is used to acquire multiple groups of detection images of each metal powder packaging bag, and the vision acquisition module transmits the detection image module to the image processing module;

[0109] After receiving the detection image, the image processing module matches the detection image with the corresponding technical powder packaging bag, and the image processing module preprocesses the detection image to improve the quality of the detection image;

[0110] After receiving the pre - processed detection images, the central processing module matches the detection images with the corresponding technical powder packaging bags. When the visual acquisition module acquires multiple detection images, the image processing module establishes an image library for the corresponding packaging bags. The image library is used to store each detection image of the corresponding packaging bag, and the database distributes each detection image in a time series;

[0111] Visually identify each detection image in the time series, and respectively obtain the usage quality data of the corresponding detection image in the packaging bag time series. The usage quality data includes the continuity value C of the packaging bag edge e and the clarity C of the image elements q and the spatial correlation value D between the image elements p and the degree of unfolding T of the packaging bag surface s ;

[0112] The calculation formula for the clarity C of the image elements q is:

[0113]

[0114] The spatial correlation value D p adopts the formula:

[0115]

[0116] The degree of unfolding T of the surface s adopts the formula:

[0117]

[0118] The central processing module removes the unit and takes the numerical value of the usage quality data of each detection image in the packaging bag time series, and calculates the quality score of the corresponding packaging bag through a formula. The formula for the quality score is:

[0119]

[0120] Finally, the central processing module uses the average value of the quality scores of each detection image in the corresponding packaging bag time series as the final quality score of the packaging bag;

[0121] Compared with Embodiment 1, the metal powder packaging quality detection system in Embodiment 2 includes a pressing module. The pressing module includes an execution unit and a pressure monitoring unit;

[0122] When the central processing module calculates the quality score Q of the corresponding packaging bag and the quality score Q ≥ the standard score, the execution unit starts, applies a preset pressure to a preset part of the packaging bag, and the central processing module monitors the pressure applied to the packaging bag by the execution unit through the pressure monitoring unit;

[0123] Subsequently, the visual acquisition module acquires the recognition image of the corresponding packaging bag. The central processing module compares the recognition image with the detection image to obtain the deformation area and deformation amplitude of the packaging bag in the recognition image. And the central processing module detects the deformed part of the packaging bag through the formula:

[0124]

[0125] In the formula, T d is the detection score of the deformed part, which is used to measure whether the position, quantity and amplitude of the inflated part of the packaging bag meet the expectations. The lower this value is, the more normal the inflation situation of the packaging bag is; the higher the value is, the more serious the inflation problem is;

[0126] d is the number of deformed parts; A i is the area of the i-th deformed part, and its area of each deformed part is calculated through region analysis; A max is the maximum allowable deformation area, which is set based on the design standard and packing standard of the corresponding packaging bag and is used to measure whether each inflated part meets the allowable range; V i is the deformation amplitude of the i-th deformed part, which is measured by calculating the relative change amount of the deformed area in the detection image and the recognition image; V max is the maximum allowable deformation amplitude, which is the upper limit of the deformation amplitude specified during the design of the packaging bag and is used to ensure that the shape of the packaging bag meets the expectations; x i and y i are the coordinates of the i-th deformed part; is the target coordinate of the i-th deformed part, which represents the standard position where the deformed part should be located and is usually set according to the packaging bag design or the target area; γ is the weight coefficient of the deformed part coordinate limit, which is used to adjust the influence degree of the offset of the inflated part from the standard position on the detection score of the deformed part;

[0127] And the central processing module compares the calculated detection score of the deformed part with the preset deformation score threshold. When the detection score of the deformed part < the deformation score threshold and the quality score Q ≥ the standard score, the central processing module marks the corresponding packaging bag as qualified.

[0128] Embodiment 3

[0129] Compared with Embodiment 2, the metal powder packaging quality detection system in Embodiment 3 includes a pressing module, and the pressing module includes an execution unit and a pressure monitoring unit;

[0130] When the central processing module calculates the quality score Q of the corresponding packaging bag and the quality score Q ≥ the standard score, the execution unit is started to apply a preset pressure to the preset part of the packaging bag, and the central processing module monitors the pressure applied to the packaging bag by the execution unit through the pressure monitoring unit;

[0131] Subsequently, the visual acquisition module acquires the recognition image of the corresponding packaging bag. The central processing module compares the recognition image with the detection image to obtain the deformation area and deformation amplitude of the packaging bag in the recognition image. Moreover, the central processing module detects the deformed part of the packaging bag through a formula, and the formula is:

[0132]

[0133] Moreover, when the central processing module calculates the recognition image, the central processing module records the position and amplitude of the deformed part of the packaging bag, and marks them as the initial position and initial amplitude respectively;

[0134] Subsequently, the execution unit continuously applies pressure to the packaging bag, the visual acquisition module continuously acquires the recognition image of the corresponding packaging bag, and the central processing module arranges the continuously acquired recognition images in a time series to obtain a set of recognition images;

[0135] The central processing module compares the position and amplitude of the deformed part of the packaging bag in the set of recognition images with the corresponding initial position and initial amplitude in a time series, and obtains the deformation position difference and deformation amplitude difference respectively;

[0136] Finally, the central processing module compares the deformation position difference and deformation amplitude difference with the preset position difference threshold and amplitude difference threshold respectively;

[0137] When the detection score of the deformed part < the deformation score threshold, the quality score Q ≥ the standard score, the deformation position difference < the position difference threshold, and the deformation amplitude difference < the amplitude difference threshold, the central processing module marks the corresponding packaging bag as qualified.

[0138] Embodiment 4

[0139] Compared with Embodiment 1, a metal powder packaging quality detection system based on machine vision in Embodiment 4 includes a visual acquisition module, an image processing module, and a central processing module;

[0140] The visual acquisition module is used to acquire at least one set of detection images of each metal powder packaging bag, and the visual acquisition module transmits the detection image module to the image processing module;

[0141] After receiving the detection image, the image processing module matches the detection image with the corresponding technical powder packaging bag, and the image processing module preprocesses the detection image to improve the quality of the detection image;

[0142] After receiving the preprocessed detection image, the central processing module performs visual recognition on the detection image to obtain the usage quality data of the packaging bag. The usage quality data includes the continuity value C of the packaging bag edge e 、the clarity C of the image elementsq The spatial correlation value D between image elements p and the degree of unfolding T of the packaging bag surface s ;

[0143] Among them, the image elements include text parts (such as production date, batch number, specification), bar code parts (such as barcodes, two-dimensional codes), and label parts;

[0144] And before the central processing module calculates the clarity C of the image elements in the corresponding packaging bag, it identifies the content of each image element and determines whether the content of the image element of the corresponding packaging bag matches the preset content of the packaging bag image element to determine whether the content of the image element of the corresponding packaging bag meets the predetermined standard or sequence; q If the match is successful, the central processing module calculates the clarity C of the image elements

[0145] ; q Calculation;

[0146] If the match fails, the packaging bag is marked as a suspicious package;

[0147] After the match is successful, the central processing module performs visual recognition on the detection image to obtain the usage quality data of the packaging bag. The usage quality data includes the continuity value C of the packaging bag edge e The clarity C of the image elements q The spatial correlation value D between image elements p and the degree of unfolding T of the packaging bag surface s ;

[0148] The clarity C of the image elements q The calculation formula is:

[0149]

[0150] The spatial correlation value D p The formula adopted is:

[0151]

[0152] The surface unfolding degree T s The formula adopted is:

[0153]

[0154] The central processing module removes the unit of the usage quality data of the packaging bag and takes its numerical value, and calculates the quality score of the corresponding packaging bag through a formula. The formula adopted for the quality score is:

[0155]

[0156] The above are only the preferred embodiments of the present invention, and do not impose any form of limitation on the present invention. Although the present invention has been disclosed above with the preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some changes or modifications to equivalent embodiments by using the above-disclosed technical content without departing from the technical solution of the present invention. However, as long as it does not depart from the technical solution content of the present invention, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention still fall within the scope of the technical solution of the present invention.

Claims

1. A metal powder packaging quality inspection system based on machine vision, including a visual acquisition module, an image processing module, and a central processing module, characterized in that: The visual acquisition module is used to acquire at least one set of detection images of each metal powder packaging bag, and the visual acquisition module transmits the detection image module to the image processing module; After receiving the detection image, the image processing module matches the detection image with the corresponding technical powder packaging bag, and the image processing module pre-processes the detection image; After receiving the pre-processed detection image, the central processing module performs visual recognition on the detection image to obtain the use quality data of the packaging bag. The use quality data includes the continuous value C of the edge of the packaging bag. e , the clarity of image elements C q , the spatial correlation value D between image elements p and the expansion degree T of the packaging bag surface s ; The central processing module removes the unit of the quality data of the packaging bag and obtains the value, and calculates the quality score of the corresponding packaging bag through a formula. The formula used for the quality score is: Where D r is the standard spatial correlation value of the image elements in the corresponding packaging bag; T r is the standard surface expansion of the corresponding packaging bag, indicating the standard value of the surface expansion of the packaging bag; e is a constant to avoid the denominator being 0; w1, w2, w3 and w4 are weight coefficients; The central processing module compares the quality score Q of the corresponding packaging bag with a preset standard score. When the quality score Q≥the standard score, the central processing module marks the corresponding packaging bag as qualified.

2. The metal powder packaging quality inspection system based on machine vision according to claim 1 is characterized in that: When the visual acquisition module acquires multiple detection images, the image processing module establishes an image library corresponding to the packaging bag, the image library is used to store each detection image of the corresponding packaging bag, and the database distributes each detection image in time series.

3. The metal powder packaging quality inspection system based on machine vision according to claim 1 is characterized in that: The central processing module calculates the continuous value C of the edge of the packaging bag e , the formula used is: Where, L c L is the length of the continuous part of the packaging bag; t is the total edge length of the packaging bag; α is the weight coefficient of the number of notches, N g is the number of edge notches; β is the weight coefficient of the edge curvature of the packaging bag; K is the edge curvature of the packaging bag.

4. The metal powder packaging quality inspection system based on machine vision according to claim 3 is characterized in that: The central processing module calculates the clarity C of the image elements q The calculation formula is: In the formula, I(x i ,y i ) is the image at position (x i ,y i ) pixel value; I(x i-1 ,y i-1 ) is the image at position (x i-1 ,y i-1 ) is the previous pixel value; n is the number of pixels in the calculation area.

5. The metal powder packaging quality inspection system based on machine vision according to claim 4 is characterized in that: The central processing module calculates the spatial correlation value D between image elements. p The formula is: In the formula, x i ,y i is the actual position coordinate of the i-th image element; is the target position coordinate of the i-th image element; m is the number of image elements.

6. The metal powder packaging quality inspection system based on machine vision according to claim 5 is characterized in that: The central processing module calculates the expansion degree T of the packaging bag surface s The formula is: Where W and H are the width and height of the image respectively; Mean(h) is the average pixel value of the grayscale values ​​of all pixels in the grayscale image after the detection image is converted into a grayscale image.

7. The metal powder packaging quality inspection system based on machine vision according to claim 4 is characterized in that: The central processing module calculates the clarity C of the image elements in the corresponding packaging bag q Before the packaging bag is packaged, the content of each image element is identified, and it is determined whether the content of the corresponding packaging bag image element matches the preset packaging bag image element content, so as to determine whether the content of the corresponding packaging bag image element meets the predetermined standard or sequence; If the match is successful, the central processing module performs the image element clarity C q Calculation.

8. The metal powder packaging quality inspection system based on machine vision according to claim 6 is characterized in that: The metal powder packaging quality detection system includes a pressing module, and the pressing module includes an execution unit and a pressure monitoring unit; When the central processing module calculates that the quality score Q of the corresponding packaging bag is ≥ the standard score, the execution unit is started to apply a preset pressure to a preset part of the packaging bag, and the central processing module monitors the pressure applied by the execution unit to the packaging bag through the pressure monitoring unit; Then the visual acquisition module acquires the recognition image of the corresponding packaging bag, and the central processing module compares the recognition image with the detection image to obtain the deformation area and deformation amplitude of the packaging bag in the recognition image; And the central processing module compares the calculated deformation part detection score with the preset deformation score threshold. When the deformation part detection score is less than the deformation score threshold, the central processing module marks the corresponding packaging bag as qualified.

9. The metal powder packaging quality inspection system based on machine vision according to claim 8, characterized in that: When the central processing module calculates the recognition image, the central processing module records the position and amplitude of the deformation part of the packaging bag, and marks them as the initial position and initial amplitude respectively; Then the execution unit continuously applies pressure to the packaging bag, the visual acquisition module continuously acquires the recognition image of the corresponding packaging bag, and the central processing module processes the continuously acquired recognition images in time series to obtain a recognition image set; The central processing module compares the position and amplitude of the deformation part of the packaging bag in the recognition image set with the corresponding initial position and initial amplitude in time series to obtain the deformation position difference and deformation amplitude difference respectively; Finally, the central processing module compares the deformation position difference and deformation amplitude difference with the preset position difference threshold and amplitude difference threshold respectively; When the deformation position difference is less than the position difference threshold, and the deformation amplitude difference is less than the amplitude difference threshold, the central processing module marks the corresponding packaging bag as qualified.

Images