CN114113119A - Product visual quality inspection method, system, equipment and medium based on defect grade - Google Patents
Product visual quality inspection method, system, equipment and medium based on defect grade Download PDFInfo
- Publication number
- CN114113119A CN114113119A CN202111466440.2A CN202111466440A CN114113119A CN 114113119 A CN114113119 A CN 114113119A CN 202111466440 A CN202111466440 A CN 202111466440A CN 114113119 A CN114113119 A CN 114113119A
- Authority
- CN
- China
- Prior art keywords
- product
- defect
- sorted
- grade
- sorting signal
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 230000007547 defect Effects 0.000 title claims abstract description 325
- 230000000007 visual effect Effects 0.000 title claims abstract description 72
- 238000007689 inspection Methods 0.000 title claims abstract description 54
- 238000000034 method Methods 0.000 title claims abstract description 45
- 238000004458 analytical method Methods 0.000 claims abstract description 106
- 238000012545 processing Methods 0.000 claims abstract description 32
- 230000002950 deficient Effects 0.000 claims description 46
- 238000004590 computer program Methods 0.000 claims description 12
- 108010076504 Protein Sorting Signals Proteins 0.000 claims description 11
- 238000004364 calculation method Methods 0.000 claims description 7
- 230000007274 generation of a signal involved in cell-cell signaling Effects 0.000 claims description 4
- 238000003860 storage Methods 0.000 claims description 3
- 238000001514 detection method Methods 0.000 description 19
- 238000004422 calculation algorithm Methods 0.000 description 10
- 238000005520 cutting process Methods 0.000 description 8
- 238000013135 deep learning Methods 0.000 description 4
- 238000004519 manufacturing process Methods 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 3
- 238000012216 screening Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005728 strengthening Methods 0.000 description 1
- 230000001960 triggered effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8854—Grading and classifying of flaws
Abstract
The invention relates to a product visual quality inspection method, a product visual quality inspection system, a product visual quality inspection device and a product visual quality inspection medium based on defect grades. A visual quality inspection method for a product based on defect grades comprises the following steps: acquiring an image of a product to be sorted to obtain a product image; performing visual analysis processing on the product image to generate a corresponding product analysis result; judging whether the product to be inspected is a good product or not according to the product analysis result, if so, generating a good product inspection signal, otherwise, continuously judging the defect grade of the product to be inspected, and generating a defect inspection signal according to the defect grade; determining a pulling distance according to the good product sorting signal and the defect sorting signal; and pulling the product to be sorted according to the pulling distance so that the product to be sorted is sorted after being pulled.
Description
Technical Field
The invention relates to the technical field of product quality detection, in particular to a product visual quality detection method, system, equipment and medium based on defect grades.
Background
For a finished zipper product, a plurality of zipper strips are combined into a long zipper by head and tail connection before entering a cutting link, a zipper head can be clamped by a clamp on a production line, and the zipper is cut off by a cutting machine after being pulled to a fixed position, so that a complete independent zipper is separated, then the clamp can move to the cutting position again, the zipper head is clamped again for pulling, and the process is repeated. In the process of clamping the zipper by the clamp hand, firstly, photographing the zipper by a camera; inputting the picture into a rear-end detection algorithm for detection while taking a picture; cutting off the zipper by a cutter; acquiring a detection result of the zipper, performing sorting on the zipper according to the detection result, namely uniformly placing the zippers with COOD detection results into qualified products through sorting equipment (such as a manipulator or a gate), uniformly placing the zippers with NG detection results into defective products, and then manually performing secondary detection on the zippers in NG to further screen out good products, wherein the rest is the real defective products.
However, in the actual detection, in order to ensure the accuracy of the GOOD product, it is preferred to ensure that all the zippers placed in the GOOD are GOOD, and all the zippers which may be defective are judged to be NG, which results in a high NG false judgment rate and a large number of GOOD products are mistaken for defective.
Because the good products in NG are more in quantity, in the actual working link, the quality inspection worker often can still screen out and put into the good product as the good product because of fatigue or in order to accelerate the working progress to produce artificial pollution, put the substandard product into the good product and send for the customer, seriously influence product quality, some can cause serious quality problems even. Therefore, how to reduce the defective products manually introduced in the secondary sorting process becomes a very critical problem in the visual quality inspection of the zipper.
Disclosure of Invention
The invention aims to overcome at least one defect in the prior art, and provides a product visual quality inspection method, a system, equipment and a medium based on defect grades, which are used for sorting products according to the defect types of the products, are beneficial to quickly screening products with different defects in secondary sorting, reduce the error rate of the secondary sorting and improve the product quality.
The invention adopts the technical scheme that a product visual quality inspection method based on defect grade comprises the following steps:
acquiring an image of a product to be sorted to obtain a product image;
performing visual analysis processing on the product image to generate a corresponding product analysis result;
judging whether the product to be inspected is a good product or not according to the product analysis result, if so, generating a good product inspection signal, otherwise, continuously judging the defect grade of the product to be inspected, and generating a defect inspection signal according to the defect grade;
determining a pulling distance according to the good product sorting signal and the defect sorting signal;
and pulling the product to be sorted according to the pulling distance so that the product to be sorted is sorted after being pulled.
The good products and the defective products with different defect levels can be sorted through the pulling distance in the primary sorting, the defective products and the defective products with different defect levels can be visually, accurately and quickly distinguished according to the pulling distance, so that the defective products are further subdivided, the error probability of the manual secondary sorting is reduced, and the delivery quality of the final products is improved.
Further, a plurality of defect types are preset, one defect type corresponds to one defect grade, one defect grade corresponds to one defect sorting signal,
the continuously judging the defect grade of the product to be sorted and generating a defect sorting signal according to the defect grade comprises the following steps:
continuously judging one or more defect types of the products to be sorted, determining one or more corresponding defect grades according to one or more defect types,
and generating a corresponding defect sorting signal according to one defect grade, or generating a defect sorting signal corresponding to the defect grade with the highest grade according to a plurality of defect grades.
Presetting a plurality of defect types of a defective product corresponding to a plurality of defect grades, continuously judging the defect type of the product to be inspected when judging that the product to be inspected is not a good product according to a visual analysis result, determining a defect grade corresponding to one defect type when the product to be inspected has one defect type, and generating a defect inspection signal corresponding to the defect grade; when a plurality of defect types simultaneously appear on a product to be sorted, determining a plurality of corresponding defect grades according to the plurality of defect types, and generating a defect sorting signal corresponding to the highest grade in the plurality of defect grades, namely the defect grade with the largest severity. The defect classification method comprises the steps of presetting a plurality of defect types corresponding to a plurality of defect grades, and determining one or more defect grades according to one or more defect types appearing in a product to be classified, so that a defect classification signal is obtained, the product to be classified is subjected to well-ordered classification, and disorder in classification is avoided.
Further, the determining a pulling distance according to the good product sorting signal and the defect sorting signal includes:
presetting that the good product sorting signal and the defect sorting signal respectively correspond to a pulling distance, and determining the pulling distance corresponding to the product to be sorted according to the good product sorting signal and the defect sorting signal.
The method for determining the pulling distance according to the good product sorting signal and the defect sorting signal comprises the steps of determining a product to be sorted to be a good product according to the good product sorting signal, determining defect products with different defect grades according to different defect sorting signals, presetting the pulling distance corresponding to the good product sorting signal and the different defect sorting signals, and determining the pulling distance corresponding to the product to be sorted according to the received good product sorting signal and the different defect sorting signals. Through the different pulling distances of presetting the defective products of yields and different defect grades, realize carrying out the go-no-go to the defective products of yields and different defect grades, further subdivide the defect product, avoid artifical secondary go-no-go mistake.
Further, the determining a pulling distance according to the good product sorting signal and the defect sorting signal includes:
determining the pulling distance of the product to be sorted to be 0 according to the good product sorting signal;
and determining the pulling distance of the product to be sorted as the corresponding defect grade multiplied by a preset pulling constant according to the defect sorting signal.
The method for determining the pulling distance according to the good product sorting signal and the defect sorting signal comprises the steps of determining that a product to be sorted is a good product according to the good product sorting signal, wherein the pulling sorting of the signal to be sorted is not needed at the moment, so that the pulling distance of the product to be sorted is determined to be 0; the defect products with different defect grades can be determined according to different defect sorting signals, a pulling constant is preset, and when the defect sorting signals corresponding to the defect grades are obtained, the pulling distance of the products to be sorted can be calculated and determined according to the defect grades and the pulling constant. The pulling distance corresponding to different defect grades is 0 through the preset good product pulling distance, the higher product of the defect grade is pulled to the farther position, products with different defect grades are distinguished according to the distance, sorting of the good product and the defect products with different defect grades is realized, the defect products are further subdivided, and errors in manual secondary sorting are avoided.
Further, the image acquisition of the product to be sorted to obtain a product image includes:
after acquiring images of a product to be inspected to obtain a plurality of product images, judging whether the acquisition of all the product images is finished, if so, performing visual analysis processing according to the plurality of product images;
the generating of the corresponding product analysis result after the visual analysis processing of the product image comprises:
and respectively carrying out visual analysis processing on the plurality of product images to generate a plurality of corresponding analysis results, and integrating the plurality of analysis results to generate a product analysis result corresponding to the product to be sorted.
In the process of product image acquisition, sometimes a plurality of product images need to be acquired to complete the analysis of a product, before visual analysis, whether the product image acquisition of a product to be sorted is completed or not is judged, after all the product image acquisition is completed, the visual analysis of the plurality of product images of the product is started, each product image is analyzed one by one, and after a plurality of corresponding analysis results are obtained, the product analysis results of the product to be sorted are integrated. The method has the advantages that complete and accurate visual analysis can be performed on one product to be sorted by adopting a plurality of product images, the accuracy of analysis results is improved, a plurality of analysis results are integrated into one product analysis result, and sorting efficiency and accuracy are improved.
The invention adopts another technical scheme that a product quality inspection system based on defect grade comprises:
the image acquisition module is used for acquiring images of products to be inspected to obtain product images;
the visual analysis module is used for carrying out visual analysis processing on the product image to generate a corresponding product analysis result;
the signal generation module is used for judging whether the product to be inspected is a good product or not according to the product analysis result, if so, generating a good product inspection signal, otherwise, continuously judging the defect grade of the product to be inspected, and generating a defect inspection signal according to the defect grade;
the distance calculation module is used for determining a pulling distance according to the good product sorting signal and the defect sorting signal;
and the sorting processing module is used for pulling the products to be sorted according to the pulling distance so as to carry out sorting processing on the products to be sorted after being pulled.
According to the invention, different pulling distances are determined through the distance calculation module, products to be inspected with different defect grades are quickly inspected, compared with the method that in the prior art, good products and defective products are firstly inspected for the first time, and then the defective products are inspected for the second time through manual work, in the primary inspection, the good products and the defective products with different defect grades can be inspected through the pulling distances, the defective products are not required to be judged manually in the subsequent secondary inspection, and the defective products with different defect degrees can be visually, accurately and quickly distinguished according to the pulling distances, so that the defective products are further subdivided, the error probability of the manual secondary inspection is reduced, and the delivery quality of the final products is improved.
Further, presetting a plurality of defect types, wherein one defect type corresponds to one defect grade, and one defect grade corresponds to one defect sorting signal;
if not, continuing to judge the defect grade of the product to be sorted, and generating a defect sorting signal according to the defect grade, wherein the defect sorting signal comprises the following steps:
if not, continuously judging one or more defect types of the products to be sorted, determining one or more defect grades according to one or more defect types,
and generating a corresponding defect sorting signal according to one defect grade, or generating a defect sorting signal corresponding to the defect grade with the highest grade according to a plurality of defect grades.
Further, the determining a pulling distance according to the good product sorting signal and the defect sorting signal includes:
presetting that the good product sorting signal and the defect sorting signal respectively correspond to a pulling distance, and determining the pulling distance corresponding to the product to be sorted according to the good product sorting signal and the defect sorting signal.
Another technical solution adopted by the present invention is a computer device, including a memory and a processor, where the memory stores a computer program, and the processor implements the steps of the product visual quality inspection method based on the defect grade when executing the computer program.
Another technical solution adopted by the present invention is a computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the steps of the product visual quality inspection method based on the defect grade.
Compared with the prior art, the invention has the beneficial effects that: the visual quality inspection method, system, equipment and medium for the products based on the defect grades quickly inspect the products to be inspected, which are different in good product and defect grades, by determining different pulling distances.
Drawings
FIG. 1 is a flowchart of a visual quality inspection method for a product based on defect levels according to the present invention.
Fig. 2 is a block diagram of a visual quality inspection system for a product based on defect levels according to the present invention.
Detailed Description
The drawings are only for purposes of illustration and are not to be construed as limiting the invention.
With the development of computer vision technology, many enterprises have started to automatically find the defects of industrial products and perform sorting by image detection methods, and the steps are generally taken as follows: firstly, photographing an industrial product; secondly, inputting the picture into a rear-end detection algorithm for detection; after the detection result is obtained, the industrial products are sorted according to the detection result, that is, the industrial products (also called as GOOD products) with the detection result of GOOD are uniformly placed in a qualified area through sorting equipment (such as a mechanical arm or a gate), and the industrial products (also called as defective products) with the detection result of NG are uniformly placed in a defective area.
The industrial products are generally classified into primary sorting and secondary sorting through sorting equipment, the primary sorting process comprises the steps of grabbing and pulling on a conveying belt or through grabbing equipment, for example, for a finished product of a zipper, a zipper head is clamped through a clamping hand and pulled to a fixed position, then the zipper head is cut off through a cutting machine, a complete independent zipper is separated, a good product is placed in a qualified area, the defective product is divided into a defective area, then the clamping head moves to the cutting position again, the zipper head is clamped again, pulling is carried out, and the process is repeated; and in the secondary sorting process, the zippers in the defect areas are subjected to secondary sorting manually, so that good products are further screened out, and the rest products are real defective products.
Aiming at the problems that secondary sorting causes inaccurate sorting of defective products and seriously affects the delivery quality of the products, the existing methods are all avoided by artificially strengthening management and requiring workers to strengthen quality consciousness. The method still cannot effectively control the sorting quality of the link, because the labor is easy to fatigue, and particularly for workers in night shifts, the error is easy to make. In order to overcome the problems in the prior art, the invention adopts a product visual quality inspection method, a system, equipment and a medium based on defect grades, and inspects products by pulling distance, thereby being beneficial to accurately and quickly screening products with different defects in secondary inspection and reducing the error rate of product inspection.
Example 1
Referring to fig. 1, fig. 1 is a flowchart illustrating a product visual quality inspection method based on defect levels according to an embodiment of the present invention, the method including:
s101, carrying out image acquisition on a product to be sorted to obtain a product image;
s102, performing visual analysis processing on the product image to generate a corresponding product analysis result;
specifically, the image acquisition of the product to be sorted to obtain the product image includes:
after acquiring images of a product to be inspected to obtain a plurality of product images, judging whether the acquisition of all the product images is finished, if so, performing visual analysis processing according to the plurality of product images;
the generating of the corresponding product analysis result after the visual analysis processing of the product image comprises:
and respectively carrying out visual analysis processing on the plurality of product images to generate a plurality of corresponding analysis results, and integrating the plurality of analysis results to generate a product analysis result corresponding to the product to be sorted.
In an actual product production line, visual analysis of a product can be realized through one or more product images, but for products which are long or have large areas, one picture cannot cover the whole product, and a plurality of pictures are required to be shot to cover a complete product, and the analysis of the plurality of pictures can improve the accuracy of product analysis results and sorting.
In this embodiment, a visual analysis algorithm is specifically adopted to perform visual analysis processing on a product image to obtain a product analysis result, the product analysis result is a good product or a defective product, the visual analysis algorithm can adopt an existing deep learning algorithm model and the like, and the deep learning algorithm model can accurately and efficiently complete detection and analysis of the product image.
In the step S101, under the condition that a plurality of product images may be acquired for the product to be sorted, a visual analysis process is performed on one product image to obtain an analysis result corresponding to one product image, after the visual analysis is sequentially performed on the plurality of product images of the product to be sorted, a plurality of analysis results of the product to be sorted are obtained, and the plurality of analysis results are integrated into one product analysis result, so that a sorting signal is subsequently generated only according to one product analysis result, a sorting process is performed, and the efficiency and accuracy of sorting are improved.
S103, judging whether the product to be inspected is a good product or not according to the product analysis result, if so, generating a good product inspection signal, otherwise, continuously judging the defect grade of the product to be inspected, and generating a defect inspection signal according to the defect grade;
specifically, a plurality of defect types are preset, one defect type corresponds to one defect grade, one defect grade corresponds to one defect sorting signal,
the continuously judging the defect grade of the product to be sorted and generating a defect sorting signal according to the defect grade comprises the following steps:
continuously judging one or more defect types of the products to be sorted, determining one or more corresponding defect grades according to one or more defect types,
and generating a corresponding defect sorting signal according to one defect grade, or generating a defect sorting signal corresponding to the defect grade with the highest grade according to a plurality of defect grades.
In this embodiment, since the product has a plurality of defect types, one defect type is preset to correspond to one defect level according to the severity of the plurality of defect types, and a defect sorting signal is generated for each defect level when the defect level corresponding to the defect type with more severe defects is higher. When the products to be sorted are judged to be good products according to the product analysis results, correspondingly generating good product sorting signals; and when the products to be sorted are judged to be defective according to the product analysis result, continuously judging the defect types of the products to be sorted, generating corresponding defect grades according to the defect types, and generating defect sorting signals according to the defect grades. It is worth noting that the products to be inspected may have one or more defect types, when one defect type occurs, a defect grade corresponding to the defect type is undoubtedly generated, and a corresponding defect inspection signal is generated according to the defect grade; when a plurality of defect types appear, a plurality of defect grades corresponding to the plurality of defect types are generated, and a defect sorting signal corresponding to the defect grade with the highest grade in the plurality of defect grades is generated according to the plurality of defect grades.
S104, determining a pulling distance according to the good product sorting signal and the defect sorting signal;
specifically, the determining a pulling distance according to the good product sorting signal and the defect sorting signal includes:
presetting that the good product sorting signal and the defect sorting signal respectively correspond to a pulling distance, and determining the pulling distance corresponding to the product to be sorted according to the good product sorting signal and the defect sorting signal.
Specifically, the determining a pulling distance according to the good product sorting signal and the defect sorting signal includes:
determining the pulling distance of the product to be sorted to be 0 according to the good product sorting signal;
and determining the pulling distance of the product to be sorted as the corresponding defect grade multiplied by a preset pulling constant according to the defect sorting signal.
S105, pulling the products to be sorted according to the pulling distance, so that the products to be sorted are sorted after being pulled.
This embodiment places the defective products of non-defective products and different defect grades in the region of difference through the pulling distance, can realize distinguishing defective products of non-defective products and different defect types, and the follow-up quality inspection workman of being convenient for is to the further subdivision screening of defect product to avoided the zip fastener too much, not had the primary and secondary, and will have the artificial error that the defective products were leaked into to the serious defect zip fastener easily.
The method for distinguishing the good products and the different defective products by determining the pulling distance includes, but is not limited to, two methods listed in this embodiment, one method is that a preset good product sorting signal and different defect sorting signals correspond to one pulling distance, that is, one pulling distance is set according to the good products and the defective products with multiple defect grades, when the products to be sorted are judged to be good products or defective products with different defect grades, the products to be sorted are cut off, and then the products to be sorted are pulled to the preset position according to the pulling distance, so that the good products or the defective products with different defect types can be placed at different positions;
secondly, presetting a pulling distance corresponding to a good product sorting signal to be 0, and setting the pulling distance corresponding to different defect sorting signals to be a defect grade multiplied by a preset pulling constant, namely, when the product to be sorted is judged to be a good product, cutting the product to be sorted, and directly sorting the product to be sorted into a good product area without pulling the product to be sorted; when the product to be inspected is judged to be a defective product, after the product to be inspected is cut, a pulling distance is calculated and determined according to the defect grade of the product to be inspected, and then the product to be inspected is placed at a corresponding position according to the pulling distance, it can be understood that, for example, the pulling constant is D, if the defect grade of the product to be inspected is a first defect grade, the pulling distance is D, if the defect grade of the product to be inspected is a second defect grade, the pulling distance is 2D, if the defect grade of the product to be inspected is a third defect grade, the pulling distance is 3D, therefore, when the defect grade of the product to be inspected is higher, the defects of the product to be inspected are more serious, the pulled distance is longer, and a quality inspection worker can further judge the severity of the defects of the product through the distance in the subsequent inspection process.
Example 2
Referring to fig. 2, fig. 2 is a structural diagram of a product visual quality inspection system based on defect levels according to the present embodiment, the system mainly includes five software modules, which can be connected to two hardware devices, respectively an industrial camera and a sorting device, wherein the system includes:
the image acquisition module 201 is used for acquiring images of products to be inspected to obtain product images;
the visual analysis module 202 is connected to the image acquisition module 201, and is configured to perform visual analysis on the product image to generate a corresponding product analysis result;
specifically, the image acquisition module 201 is further configured to: after acquiring images of a product to be inspected to obtain a plurality of product images, judging whether the acquisition of all the product images is finished, if so, performing visual analysis processing according to the plurality of product images;
the vision analysis module 202 is further configured to: and respectively carrying out visual analysis processing on the plurality of product images to generate a plurality of corresponding analysis results, and integrating the plurality of analysis results to generate a product analysis result corresponding to the product to be sorted.
In this embodiment, the image obtaining module 201 is connected to one or more industrial cameras, and is configured to obtain a product image captured by the industrial cameras, and send the obtained product image to the visual analysis module 202, so that the visual analysis module 202 performs visual analysis processing on the product image; when a plurality of images are required to be shot for a product to cover the complete product, the image acquisition module 201 is further configured to determine whether to complete the acquisition of all product images of the product to be sorted, and when it is determined that the acquisition of the product images is completed, the visual analysis module 202 is triggered to perform visual analysis processing on the product images.
The visual analysis module 202 specifically performs visual analysis processing on the product image by using a visual analysis algorithm to obtain a product analysis result, where the product analysis result is a good product or a defective product, the visual analysis algorithm may use an existing deep learning algorithm model, and the deep learning algorithm model can accurately and efficiently complete detection and analysis of the product image.
The visual analysis module 202 obtains the product image sent by the image obtaining module 201, and performs visual analysis processing on the product image to obtain a product analysis result, when a plurality of product images of a product to be checked need to be subjected to visual analysis processing, the visual analysis module 202 performs visual analysis processing on one product image to obtain an analysis result corresponding to one product image, after the plurality of product images of the product to be checked are subjected to visual analysis, a plurality of analysis results of the product to be checked are obtained, the plurality of analysis results are integrated into one product analysis result, and one product analysis result is sent to the signal generating module 203, so that the signal generating module 203 executes subsequent processing after obtaining the product analysis result.
A signal generating module 203, configured to determine whether the product to be sorted is a good product according to the product analysis result, if so, generate a good product sorting signal, otherwise, continue to determine a defect level of the product to be sorted, and generate a defect sorting signal according to the defect level;
specifically, a plurality of defect types are preset, one defect type corresponds to one defect grade, one defect grade corresponds to one defect sorting signal,
in the signal generating module 203, continuously judging the defect level of the product to be sorted, and generating a defect sorting signal according to the defect level, including: continuously judging one or more defect types of the products to be sorted, determining one or more corresponding defect grades according to one or more defect types,
and generating a corresponding defect sorting signal according to one defect grade, or generating a defect sorting signal corresponding to the defect grade with the highest grade according to a plurality of defect grades.
In this embodiment, since the product has a plurality of defect types, one defect type is preset to correspond to one defect level according to the severity of the plurality of defect types, and a defect sorting signal is generated for each defect level when the defect level corresponding to the defect type with more severe defects is higher. When the products to be sorted are judged to be good products according to the product analysis results, the signal generation module 203 correspondingly generates good product sorting signals; when the products to be sorted are judged to be not good products according to the product analysis result, the signal generation module 203 continues to judge the defect types of the products to be sorted, generates corresponding defect grades according to the defect types, and generates defect sorting signals according to the defect grades. It is worth noting that the products to be inspected may have one or more defect types, when one defect type occurs, a defect grade corresponding to the defect type is undoubtedly generated, and a corresponding defect inspection signal is generated according to the defect grade; when a plurality of defect types appear, a plurality of defect grades corresponding to the plurality of defect types are generated, and a defect sorting signal corresponding to the defect grade with the highest grade in the plurality of defect grades is generated according to the plurality of defect grades.
A distance calculating module 204, configured to determine a pulling distance according to the good product sorting signal and the defect sorting signal;
specifically, the distance calculation module 204 specifically includes:
presetting that the good product sorting signal and the defect sorting signal respectively correspond to a pulling distance, and determining the pulling distance corresponding to the product to be sorted according to the good product sorting signal and the defect sorting signal.
Specifically, the distance calculation module 204 specifically includes:
determining the pulling distance of the product to be sorted to be 0 according to the good product sorting signal;
and determining the pulling distance of the product to be sorted as the corresponding defect grade multiplied by a preset pulling constant according to the defect sorting signal.
And the sorting processing module 205 is configured to pull the product to be sorted according to the pulling distance, so that the product to be sorted is sorted after being pulled.
The distance calculating module 204 of this embodiment may specifically use a PLC controller to determine the pulling distance corresponding to the product to be sorted, the sorting processing module 205 is connected to a sorting device in the production line, and the sorting device may be set according to actual production, for example, a cutter or the like. In the embodiment, the distance calculation module 204 and the sorting processing module 205 are used for pulling good products and defective products with different defect grades to different areas, so that the good products and the defective products with different defect types can be distinguished, and subsequent quality inspection workers can further subdivide and screen the defective products conveniently, thereby avoiding the condition that zippers are too many and have no primary and secondary defects, and easily causing human errors that the zippers with serious defects leak into the good products.
There are various methods for distinguishing good products and different defective products by determining a pulling distance through the distance calculation module 204, including but not limited to two methods listed in this embodiment, one of which is that a preset good product sorting signal and different defect sorting signals correspond to one pulling distance, that is, a pulling distance is set according to the good products and the defective products of various defect grades, when the products to be sorted are determined to be good products or defective products of which defect grade is obtained, the products to be sorted are cut off, and then the products to be sorted are pulled to a preset position according to the pulling distance, so that the good products or the defective products of different defect types can be placed at different positions;
secondly, presetting a pulling distance corresponding to a good product sorting signal to be 0, and setting the pulling distance corresponding to different defect sorting signals to be a defect grade multiplied by a preset pulling constant, namely, when the product to be sorted is judged to be a good product, cutting the product to be sorted, and directly sorting the product to be sorted into a good product area without pulling the product to be sorted; when the product to be inspected is judged to be a defective product, after the product to be inspected is cut, a pulling distance is calculated and determined according to the defect grade of the product to be inspected, and then the product to be inspected is placed at a corresponding position according to the pulling distance, it can be understood that, for example, the pulling constant is D, if the defect grade of the product to be inspected is a first defect grade, the pulling distance is D, if the defect grade of the product to be inspected is a second defect grade, the pulling distance is 2D, if the defect grade of the product to be inspected is a third defect grade, the pulling distance is 3D, therefore, when the defect grade of the product to be inspected is higher, the defects of the product to be inspected are more serious, the pulled distance is longer, and a quality inspection worker can further judge the severity of the defects of the product through the distance in the subsequent inspection process.
In a preferred embodiment, the computer device includes a memory and a processor, where the memory stores a computer program, and the processor implements the steps of the product visual quality inspection method based on the defect level when executing the computer program.
The embodiment of the present invention further provides a preferred implementation manner, that is, a computer-readable storage medium storing a computer program, where the computer program, when executed by a processor, implements the steps of the product visual quality inspection method based on the defect grade.
It should be understood that the above-mentioned embodiments of the present invention are only examples for clearly illustrating the technical solutions of the present invention, and are not intended to limit the specific embodiments of the present invention. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention claims should be included in the protection scope of the present invention claims.
Claims (10)
1. A visual quality inspection method for a product based on a defect grade is characterized by comprising the following steps:
acquiring an image of a product to be sorted to obtain a product image;
performing visual analysis processing on the product image to generate a corresponding product analysis result;
judging whether the product to be inspected is a good product or not according to the product analysis result, if so, generating a good product inspection signal, otherwise, continuously judging the defect grade of the product to be inspected, and generating a defect inspection signal according to the defect grade;
determining a pulling distance according to the good product sorting signal and the defect sorting signal;
and pulling the product to be sorted according to the pulling distance so that the product to be sorted is sorted after being pulled.
2. The visual quality inspection method of a product based on defect grade according to claim 1, characterized in that a plurality of defect types are preset, one defect type corresponding to one defect grade, and one defect grade corresponding to one defect sorting signal;
the continuously judging the defect grade of the product to be sorted and generating a defect sorting signal according to the defect grade comprises the following steps:
continuously judging one or more defect types of the products to be sorted, determining one or more corresponding defect grades according to one or more defect types,
and generating a corresponding defect sorting signal according to one defect grade, or generating a defect sorting signal corresponding to the defect grade with the highest grade according to a plurality of defect grades.
3. The method of claim 2, wherein the determining a pulling distance according to the good product sorting signal and the defect sorting signal comprises:
presetting that the good product sorting signal and the defect sorting signal respectively correspond to a pulling distance, and determining the pulling distance corresponding to the product to be sorted according to the good product sorting signal and the defect sorting signal.
4. The method of claim 2, wherein the determining a pulling distance according to the good product sorting signal and the defect sorting signal comprises:
determining the pulling distance of the product to be sorted to be 0 according to the good product sorting signal;
and calculating the pulling distance of the product to be sorted according to the defect sorting signal, wherein the pulling distance is the product obtained by multiplying the corresponding defect grade by a preset pulling constant.
5. The visual quality inspection method for the product based on the defect grade according to claim 1, wherein the image acquisition of the product to be classified to obtain the product image comprises:
after acquiring images of a product to be inspected to obtain a plurality of product images, judging whether the acquisition of all the product images is finished, if so, performing visual analysis processing according to the plurality of product images;
the generating of the corresponding product analysis result after the visual analysis processing of the product image comprises:
and respectively carrying out visual analysis processing on the plurality of product images to generate a plurality of corresponding analysis results, and integrating the plurality of analysis results to generate a product analysis result corresponding to the product to be sorted.
6. A defect-level-based product quality inspection system, comprising:
the image acquisition module is used for acquiring images of products to be inspected to obtain product images;
the visual analysis module is used for carrying out visual analysis processing on the product image to generate a corresponding product analysis result;
the signal generation module is used for judging whether the product to be inspected is a good product or not according to the product analysis result, if so, generating a good product inspection signal, otherwise, continuously judging the defect grade of the product to be inspected, and generating a defect inspection signal according to the defect grade;
the distance calculation module is used for determining a pulling distance according to the good product sorting signal and the defect sorting signal;
and the sorting processing module is used for pulling the product to be checked according to the pulling distance so that the product to be checked is subjected to sorting processing after being pulled.
7. The system of claim 6, wherein a plurality of defect types are preset, one defect type corresponding to each defect level, and one defect level corresponding to each defect sorting signal;
the continuously judging the defect grade of the product to be sorted and generating a defect sorting signal according to the defect grade comprises the following steps:
continuing to judge one or more defect types of the products to be sorted, determining one or more defect grades according to one or more defect types,
and generating a corresponding defect sorting signal according to one defect grade, or generating a defect sorting signal corresponding to the defect grade with the highest grade according to a plurality of defect grades.
8. The system of claim 7, wherein the determining a pull distance according to the non-defective product sorting signal and the defect sorting signal comprises:
determining the pulling distance of the product to be sorted to be 0 according to the good product sorting signal;
presetting a pulling distance corresponding to one defect sorting signal, and determining the pulling distance corresponding to the product to be sorted according to the defect sorting signal.
9. A computer device comprising a memory and a processor, the memory storing a computer program, wherein the processor when executing the computer program performs the steps of the method for visual quality inspection of a product based on defect levels according to any one of claims 1 to 5.
10. A computer-readable storage medium, storing a computer program, wherein the computer program, when executed by a processor, performs the steps of the method for visual quality inspection of a product based on defect levels according to any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111466440.2A CN114113119A (en) | 2021-12-03 | 2021-12-03 | Product visual quality inspection method, system, equipment and medium based on defect grade |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111466440.2A CN114113119A (en) | 2021-12-03 | 2021-12-03 | Product visual quality inspection method, system, equipment and medium based on defect grade |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114113119A true CN114113119A (en) | 2022-03-01 |
Family
ID=80366161
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111466440.2A Pending CN114113119A (en) | 2021-12-03 | 2021-12-03 | Product visual quality inspection method, system, equipment and medium based on defect grade |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114113119A (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109187567A (en) * | 2018-08-07 | 2019-01-11 | 华侨大学 | A kind of composite polycrystal-diamond surface white-spot defects detection system |
| WO2019107614A1 (en) * | 2017-11-30 | 2019-06-06 | 전자부품연구원 | Machine vision-based quality inspection method and system utilizing deep learning in manufacturing process |
| CN109926341A (en) * | 2019-03-26 | 2019-06-25 | 武汉大学 | Silicon photrouics scanning vortex thermal imaging detection platform and defect classification method |
| CN110378900A (en) * | 2019-08-01 | 2019-10-25 | 北京迈格威科技有限公司 | The detection method of product defects, apparatus and system |
| CN111299166A (en) * | 2020-01-21 | 2020-06-19 | 山东浪潮通软信息科技有限公司 | Zipper quality detection system and method based on cognitive vision |
| CN111681220A (en) * | 2020-06-04 | 2020-09-18 | 阿丘机器人科技(苏州)有限公司 | Defect detection model construction method, device and system and storage medium |
| CN213737330U (en) * | 2020-10-28 | 2021-07-20 | 迩言(上海)科技有限公司 | Multi-type defect sorting equipment for imperfect grains |
| CN113487607A (en) * | 2021-09-06 | 2021-10-08 | 深圳新视智科技术有限公司 | Defect detection method and device based on multi-view-field image |
-
2021
- 2021-12-03 CN CN202111466440.2A patent/CN114113119A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2019107614A1 (en) * | 2017-11-30 | 2019-06-06 | 전자부품연구원 | Machine vision-based quality inspection method and system utilizing deep learning in manufacturing process |
| CN109187567A (en) * | 2018-08-07 | 2019-01-11 | 华侨大学 | A kind of composite polycrystal-diamond surface white-spot defects detection system |
| CN109926341A (en) * | 2019-03-26 | 2019-06-25 | 武汉大学 | Silicon photrouics scanning vortex thermal imaging detection platform and defect classification method |
| US20200313612A1 (en) * | 2019-03-26 | 2020-10-01 | Wuhan University | Silicon photovoltaic cell scanning eddy current thermography detection platform and defect classification method |
| CN110378900A (en) * | 2019-08-01 | 2019-10-25 | 北京迈格威科技有限公司 | The detection method of product defects, apparatus and system |
| CN111299166A (en) * | 2020-01-21 | 2020-06-19 | 山东浪潮通软信息科技有限公司 | Zipper quality detection system and method based on cognitive vision |
| CN111681220A (en) * | 2020-06-04 | 2020-09-18 | 阿丘机器人科技(苏州)有限公司 | Defect detection model construction method, device and system and storage medium |
| CN213737330U (en) * | 2020-10-28 | 2021-07-20 | 迩言(上海)科技有限公司 | Multi-type defect sorting equipment for imperfect grains |
| CN113487607A (en) * | 2021-09-06 | 2021-10-08 | 深圳新视智科技术有限公司 | Defect detection method and device based on multi-view-field image |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104063873B (en) | A kind of Model For The Bush-axle Type Parts surface defect online test method based on compressed sensing | |
| CN109499908A (en) | A kind of vision detection system for machine components | |
| CN107966454A (en) | A kind of end plug defect detecting device and detection method based on FPGA | |
| KR102168724B1 (en) | Method And Apparatus for Discriminating Normal and Abnormal by using Vision Inspection | |
| CN106706656A (en) | Machine vision-based zipper detection device and method | |
| CN105478363A (en) | Defective product detection and classification method and system based on three-dimensional figures | |
| CN110146513A (en) | Method of determining defects and determining defects device | |
| CN112150439A (en) | Automatic sorting equipment and sorting method for injection molding parts | |
| CN114029237A (en) | Real-time sorting system and method for surface defects of aluminum profiles | |
| CN110020691A (en) | LCD screen defect inspection method based on the training of convolutional neural networks confrontation type | |
| CN114529510B (en) | Automatic detection and classification method for cathode copper on-line quality | |
| CN207081666U (en) | A kind of zipper detecting device based on machine vision | |
| CN116465315A (en) | Automatic screen quality detection method and system | |
| CN109712115B (en) | Automatic PCB detection method and system | |
| JP6795562B2 (en) | Inspection equipment and machine learning method | |
| CN114280073A (en) | Tile defect detection device adopting laser and camera composite detection mode and classification method | |
| CN116165220B (en) | AOI (automated optical inspection) internal view detection device and method based on artificial intelligence | |
| CN112964732A (en) | Spinning cake defect visual detection system and method based on deep learning | |
| CN112712514A (en) | Product detection system and auxiliary detection method | |
| CN114113119A (en) | Product visual quality inspection method, system, equipment and medium based on defect grade | |
| CN111486790A (en) | Full-size detection method and device for battery | |
| CN113902728A (en) | System for detecting product quality | |
| WO2017024065A1 (en) | Range-based real-time scanning electron microscope non-visual binner | |
| CN108810784A (en) | The quality detecting system of receiver, method and device in a kind of verbal system | |
| KR20230116537A (en) | Defective product sorting device and sorting method based on deep learning technology |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |