CN117237340B - Method and system for detecting appearance of mobile phone shell based on artificial intelligence - Google Patents
Method and system for detecting appearance of mobile phone shell based on artificial intelligence Download PDFInfo
- Publication number
- CN117237340B CN117237340B CN202311496504.2A CN202311496504A CN117237340B CN 117237340 B CN117237340 B CN 117237340B CN 202311496504 A CN202311496504 A CN 202311496504A CN 117237340 B CN117237340 B CN 117237340B
- Authority
- CN
- China
- Prior art keywords
- image
- area
- shell
- mobile phone
- gray
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 37
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 20
- 238000001514 detection method Methods 0.000 claims abstract description 103
- 230000007547 defect Effects 0.000 claims abstract description 49
- 238000012545 processing Methods 0.000 claims abstract description 18
- 238000012216 screening Methods 0.000 claims abstract description 14
- 238000004458 analytical method Methods 0.000 claims abstract description 9
- 238000003709 image segmentation Methods 0.000 claims abstract description 5
- 238000007781 pre-processing Methods 0.000 claims abstract description 4
- 238000012549 training Methods 0.000 claims description 35
- 239000011159 matrix material Substances 0.000 claims description 21
- 230000011218 segmentation Effects 0.000 claims description 12
- 238000004422 calculation algorithm Methods 0.000 claims description 7
- 238000013507 mapping Methods 0.000 claims description 6
- 238000007619 statistical method Methods 0.000 claims description 4
- 238000004519 manufacturing process Methods 0.000 description 10
- 238000004364 calculation method Methods 0.000 description 4
- 238000001914 filtration Methods 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 238000005286 illumination Methods 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000003708 edge detection Methods 0.000 description 1
- 238000005429 filling process Methods 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
Landscapes
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a method and a system for detecting the appearance of a mobile phone shell based on artificial intelligence. An artificial intelligence based cell phone case appearance detection system, comprising: the mobile phone shell scanning module scans the mobile phone shell to be detected to generate a shell image to be detected; the image preprocessing module is used for carrying out gray processing on the shell image to be detected to obtain a characteristic gray image; the image segmentation module is used for dividing the characteristic gray image into a plurality of image areas; the feature analysis module is used for calculating the texture feature value of each image area of the feature gray level image; the region screening module is used for screening the image region; and the appearance detection module is used for detecting appearance defects of the screened image areas. According to the invention, the texture characteristic values are obtained by analyzing the texture information of the image areas, and the proper detection model is selected according to the texture characteristic values to carry out appearance detection, so that the detection speed and detection precision of the mobile phone shell with the interference lines are improved.
Description
Technical Field
The invention relates to the technical field of defect detection, in particular to a method and a system for detecting the appearance of a mobile phone shell based on artificial intelligence.
Background
The mobile phone shell is one of important components of the mobile phone, and the production of the mobile phone shell needs to be subjected to various production processes, and defects are inevitably generated in the mobile phone shell in the process, so that the appearance of the mobile phone shell is not in line with the requirements of consumers. In order to improve the quality of the mobile phone, it is an indispensable link to perform shell detection on the mobile phone shell.
The traditional detection of the appearance of the mobile phone shell is mostly manual detection, and has the defects of low efficiency and low precision, and the detection method based on machine vision has higher precision and higher efficiency. However, in order to improve the ornamental value and the hand feeling of the mobile phone shell, more and more ornamental lines and frosted stripes are added to the mobile phone shell, so that the quality of the mobile phone is improved, and meanwhile, the appearance detection is more challenged, and the detection precision of defects of the mobile phone shell is influenced by some irregular lines.
Disclosure of Invention
In order to solve the technical problems, the invention aims to provide a method and a system for detecting the appearance of a mobile phone shell based on artificial intelligence, and the adopted technical scheme is as follows:
in a first aspect, the present invention provides a method for detecting the appearance of a mobile phone shell based on artificial intelligence, including:
scanning a mobile phone shell to be detected to generate a shell image to be detected, and carrying out graying treatment on the shell image to be detected to obtain a characteristic gray image;
obtaining a standard gray level image of a sample mobile phone shell, wherein the sample mobile phone shell does not have appearance defects, and carrying out region segmentation on the characteristic gray level image based on a preset segmentation size to obtain a plurality of image regions, so as to construct an image region set Z;
traversing each image area in the image area set Z, calculating a texture characteristic value of each image area, and for any image area, marking the image area as a first type area if the texture characteristic value is larger than a preset classification threshold value, otherwise marking the image area as a second type area, and constructing to obtain a first type area set C1 and a second type area set C2;
dividing the characteristic gray image into a plurality of image areas based on the preset dividing size, mapping each image area in the first type area set C1 and the second type area set C2 into the characteristic gray image, and determining the area type of each image area in the characteristic gray image;
and for the characteristic gray level image, sending the image area belonging to the first type area to a trained first detection model for defect detection, calculating the texture characteristic value of each image area for the image area belonging to the second type area, respectively calculating the texture difference value of the image area belonging to the second type area in the characteristic gray level image and the image area at the same position in the standard gray level image through the texture characteristic value, and sending the image area with the texture difference value larger than a preset difference threshold value in the characteristic gray level image to a second detection model for defect detection to obtain the detection result of the mobile phone shell to be detected.
Further, for the characteristic gray-scale image and the standard gray-scale image, further comprising:
the characteristic gray level image and the standard gray level image are images after filling the hole area and the logo area through mask operation;
scanning the sample mobile phone shell and performing image graying treatment to obtain a sample shell gray image, performing background segmentation treatment and connected domain analysis on the sample shell gray image, determining a plurality of connected domains of the sample shell gray image, screening the connected domains according to a preset size threshold, determining a hole area and a logo area of the sample mobile phone shell, performing filling treatment on the hole area and the logo area on the sample shell gray image through mask operation to obtain the standard gray image, and performing graying treatment on the shell image to be detected to obtain the shell gray image to be detected;
mapping the hole area and logo area of the sample mobile phone shell into the gray level image of the shell to be detected, and replacing gray level values of the hole area and logo area of the gray level image of the shell to be detected with filling content of the hole area and logo area of the sample mobile phone shell to obtain the characteristic gray level image.
Further, the calculating the texture feature value of each image area specifically includes:
for any image area, processing the image area based on a circular LBP operator to obtain an LBP characteristic image, taking an LBP value of each pixel point in the LBP characteristic image as a gray level value, determining a gray level co-occurrence matrix of the LBP characteristic image, extracting texture features of the image area through the gray level co-occurrence matrix, carrying out statistical analysis on the gray level co-occurrence matrix, calculating to obtain entropy of the gray level co-occurrence matrix, and taking the entropy of the gray level co-occurrence matrix as the texture feature value of the image area.
Further, for the first detection model and the second detection model, the training steps are as follows:
the method comprises the steps of obtaining a training sample image set, wherein the training sample image set comprises a plurality of images which have appearance defects and belong to a first type region and a plurality of images which have appearance defects and belong to a second type region, respectively carrying out defect marking on each image in the training sample image set, classifying the marked training sample images according to the type of the region, constructing a first type training data set and a second type training data set, inputting the first type training data set into a first detection model which is constructed in advance, inputting the second type training data set into a second detection model which is constructed in advance, using cross entropy as a loss function in the training process, and completing training on the first detection model and the second detection model after the loss function reaches a preset convergence condition.
Further, for the housing image to be detected, further comprising:
and after the shell image to be detected is subjected to gray processing, carrying out noise reduction processing on the shell image to be detected after gray processing through a Gaussian filter algorithm.
As another aspect of the present application, there is provided an artificial intelligence based appearance detection system for implementing the above method for detecting appearance of a mobile phone casing, including:
the mobile phone shell scanning module scans the mobile phone shell to be detected to generate a shell image to be detected;
the image preprocessing module is used for carrying out gray processing on the shell image to be detected to obtain a characteristic gray image;
the image segmentation module is used for dividing the characteristic gray image into a plurality of image areas;
the feature analysis module is used for calculating the texture feature value of each image area of the feature gray level image;
the region screening module is used for screening a plurality of image regions of the characteristic gray level image;
and the appearance detection module is used for detecting appearance defects of the image areas obtained by the area screening module to obtain detection results of the mobile phone shell to be detected.
Further, for the appearance detection module, the method further includes:
and the appearance detection module detects appearance defects of the mobile phone shell to be detected through the first detection model and the second detection model.
The beneficial effects of the invention are as follows:
according to the mobile phone shell appearance detection method and system based on the artificial intelligence, the mobile phone shell image to be detected is divided into a plurality of image areas, the texture information of each image area is analyzed, the texture characteristic value is obtained through calculation, the plurality of image areas are divided into different types according to the texture characteristic value, appearance defect detection is conducted on the image areas of different area types through the trained first detection model and the trained second detection model, the method and system are suitable for appearance detection of the mobile phone shell with interference lines, and detection speed and detection accuracy are improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to the structures shown in these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flow chart of an artificial intelligence based method for detecting the appearance of a mobile phone shell according to an embodiment of the invention.
Fig. 2 is a schematic structural diagram of an artificial intelligence based mobile phone case appearance detection system according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, some embodiments of the present application will be described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the present application. However, those of ordinary skill in the art will understand that in the various embodiments of the present application, numerous technical details have been set forth in order to provide a better understanding of the present application. However, the technical solutions claimed in the present application can be implemented without these technical details and with various changes and modifications based on the following embodiments.
The mobile phone shell appearance detection method based on the artificial intelligence is particularly suitable for defect detection of the mobile phone shell, production of the mobile phone shell needs to be subjected to multiple procedures, and complicated production procedures bring more defects while improving the appearance quality of the mobile phone. However, for higher product quality, some handset manufacturers add frosted stripes and personalized pattern lines on the initial handset housing, so that the handset has high ornamental value and unique hand feeling, and the production process of the handset housing is more and more complex, and the new process may cause appearance defects of the handset housing, and during appearance detection of the handset housing, the unique lines will interfere with defect detection, such as scratch defect detection. Under such a background, the embodiment of the invention provides a method and a system for detecting the appearance of a mobile phone shell based on artificial intelligence, which are used for solving the technical problems existing in the prior art, and the technical scheme of the invention is specifically described below by a specific implementation manner and with reference to the accompanying drawings.
Referring to fig. 1, a flow chart of an artificial intelligence based method for detecting the appearance of a mobile phone shell according to an embodiment of the present invention is provided, as an aspect of the present invention, as a method for detecting the appearance of a mobile phone shell based on an artificial intelligence, the method includes:
s100, scanning a mobile phone shell to be detected to generate a shell image to be detected, and carrying out graying treatment on the shell image to be detected to obtain a characteristic gray image;
specifically, a to-be-detected mobile phone shell irradiated by an LED light source can be scanned through a CMOS camera or a CCD camera to obtain an to-be-detected shell image, the to-be-detected shell image is subjected to gray processing, and meanwhile camera holes, logo patterns and the like in the mobile phone shell are removed to obtain a characteristic gray image.
In one embodiment, the feature gray scale image is generated by:
the method comprises the steps of scanning a sample mobile phone shell and performing image graying treatment to obtain a sample shell gray image, wherein the sample mobile phone shell is a standard mobile phone shell without appearance defects;
performing background segmentation processing and connected domain analysis on the gray level image of the sample shell, wherein the background segmentation processing and the connected domain analysis are performed on the gray level image of the sample shell by adopting an image segmentation algorithm based on edge detection, in the embodiment, the background segmentation is performed on the gray level image of the sample shell by a Canny operator, the binarization processing is performed on the image after the background segmentation to obtain a binary image, a hole area for installing a camera module, a flash lamp and the like, a logo pattern for representing mobile phone product information and the like exist in the mobile phone shell, the binary image is analyzed by the connected domain analysis algorithm, and a plurality of connected domains and the size information of each connected domain are determined;
screening the plurality of connected domains according to a preset size threshold value, and determining a hole area and a logo area of the sample mobile phone shell, wherein the preset size threshold value can be determined according to priori knowledge, and the preset size threshold value is reasonably set based on size specification information of the mobile phone shell;
performing filling treatment on the hole area and the logo area of the gray level image of the sample shell based on the pixel points in the neighborhood of the hole area and the logo area respectively by using mask operation to obtain a standard gray level image;
carrying out graying treatment on the shell image to be detected to obtain a shell gray image to be detected, mapping a hole area and a logo area of a sample mobile phone shell into the shell gray image to be detected, determining positions of the hole area and the logo area in the shell gray image to be detected, discarding gray values of pixel points of the hole area and the logo area in the shell gray image to be detected, and replacing gray values of the hole area and the logo area of the shell gray image to be detected with filling contents of the hole area and the logo area of the sample mobile phone shell to obtain a characteristic gray image;
specifically, in the process of filling the hole area and the logo area of the sample mobile phone shell through mask operation, the acquired information is respectively the gray value information of the pixels in the neighborhood of the hole area and the logo area in the gray level image of the sample mobile phone shell, and the pixels in the neighborhood of the hole area and the logo area in the gray level image of the shell to be detected possibly have defects, so that the gray value information contained in the pixel is not adopted in the filling process, and the filling information in the standard gray level image is selected to be used for replacing the gray value information.
In the process of scanning the to-be-detected mobile phone shell, the acquired to-be-detected shell image has noise due to the influences of the structures of the scanning equipment and the sensor equipment and the working environment, noise reduction treatment is needed to be carried out on the to-be-detected shell image, for example, a median filtering algorithm, a Gaussian filtering algorithm and the like are carried out on the to-be-detected shell image, and the Gaussian filtering algorithm is adopted to reduce the noise in the image in the embodiment.
S200, acquiring a standard gray level image of a sample mobile phone shell, carrying out region segmentation on a characteristic gray level image based on a preset segmentation size to obtain a plurality of image regions, and constructing an image region set Z;
specifically, the preset division size can be set according to the size information of the standard gray image, the standard gray image is divided into a plurality of image areas with the same size through the preset division size, an image area set Z is constructed, and the plurality of image areas obtained through division are recorded in the image area set Z.
S300, traversing each image area in the image area set Z, calculating a texture characteristic value of each image area, classifying each image area according to the texture characteristic value, and constructing a first type area set C1 and a second type area set C2;
specifically, after the texture feature value of each image area is obtained through calculation, for any image area, if the texture feature value is larger than a preset classification threshold value, the image area is marked as a first type area, otherwise, the image area is marked as a second type area, an image area set Z is traversed, the area type of each image area in the image area set Z is determined, the image areas belonging to the same area type are stored in the same set, and a first type area set C1 and a second type area set C2 are constructed.
In an alternative embodiment, the texture feature values are calculated as follows:
for any image area, processing the image area based on a circular LBP operator to obtain an LBP characteristic image;
the round LBP operator has gray scale invariance and rotation invariance, is affected by illumination in the process of collecting images through a CMOS camera or a CCD camera, more noise can exist in the collected images, and the characteristic information collected through the round LBP operator is little affected by illumination.
In this embodiment, the sampling radius of the circular LBP operator is 1, the sampling number is 8, the acquired LBP value of each pixel point is an eight-bit binary number, each binary number is circularly shifted to the right, multiple groups of binary numbers are converted into decimal numbers, the minimum value of the decimal numbers in the process of circularly shifting to the right is selected as the LBP value of the pixel point, the original value range of the LBP value of the eight-bit binary number is [0,255], and the value range of the LBP value after circularly shifting to the right to the minimum value is compressed to [0,36], so that the calculation amount is greatly reduced.
Taking the LBP value of each pixel point in the LBP characteristic image as a gray value, determining a gray level co-occurrence matrix of the LBP characteristic image, and extracting texture features of an image area through the gray level co-occurrence matrix;
the gray level co-occurrence matrix contains the texture information of the image, and the embodiment of the invention obtains the higher-order texture information of the image by carrying out statistical analysis on the gray level co-occurrence matrix, specifically, the entropy of the gray level co-occurrence matrix is used as the texture characteristic value of the image area by calculating the entropy of the gray level co-occurrence matrix.
For image areas belonging to different area types, the more texture information is contained, the larger the corresponding texture characteristic value is.
S400, dividing the characteristic gray image into a plurality of image areas based on a preset dividing size, mapping each image area in the first type area set C1 and the second type area set C2 into the characteristic gray image, and determining the area type of each image area in the characteristic gray image;
specifically, the feature gray image includes an indefinite number of defect information, so that when the feature gray image is divided into a plurality of image areas by a preset division size, the feature gray image is mapped into the feature gray image directly according to the previously determined classification information, namely, the first type area set C1 and the second type area set C2, so as to complete the area division of the feature gray image and the classification of each image area.
S500, performing defect detection on the image area belonging to the first type of area through a first detection model, and performing defect detection on the image area belonging to the second type of area through a second detection model to obtain a detection result of the mobile phone shell to be detected;
specifically, for the existence of ornamental lines and frosted stripes, the line information added by the production process will affect the appearance detection, such as the textures existing in the mobile phone shell and the scratch defects in the production defects, in this case, the invention divides the characteristic gray level image into a plurality of image areas and classifies each image area. The first type region does not have texture information added by a production process, and an image region belonging to the first type region is sent to a trained first detection model for defect detection; the second type of region has texture information added by the production process, for each image region belonging to the second type of region, calculating the texture feature value of each image region, comparing the image region belonging to the second type of region in the feature gray image with the image region belonging to the second type of region divided in the image region without defects such as standard gray image, if part of production defects such as scratch defects and scratch defects exist, the texture feature value of each image region is calculated, and for each image region, the texture difference value of each image region is determined by comparing the texture feature value of the image region at the same position in the standard gray image:
Tedif=|Te1-Te2|
wherein Tedif is texture difference value, te1 is texture characteristic value of image area in the characteristic gray scale image, te2 is texture characteristic value of image area in the standard gray scale image.
The texture variation of each image area is represented by the texture difference value, the larger the texture difference value is, the larger the probability of appearance defects is, the plurality of image areas are screened by the preset difference threshold, the image areas with the texture difference value larger than the preset difference threshold in the characteristic gray level image are sent to the second detection model for defect detection, and the output information of the first detection model and the second detection model is summarized, so that the detection result of the mobile phone shell to be detected is obtained.
In an alternative embodiment, for the first detection model and the second detection model, the training steps are included as follows:
obtaining a training sample image set, wherein the training sample image set comprises a plurality of images with appearance defects belonging to a first type region and a plurality of images with appearance defects belonging to a second type region, and specifically, the sources of the sample image set can be a plurality of mobile phone shells with appearance defects found in a manual detection process;
performing defect marking on each image in the training sample image set in a manual marking mode, classifying and storing marked training sample images according to the type of the region to which the images belong, constructing a first training data set and a second training data set, and respectively storing the images belonging to the first region and the second region;
the method comprises the steps of inputting a first class training data set into a first detection model which is built in advance, inputting a second class training data set into a second detection model which is built in advance, wherein the first detection model and the second detection model are both neural network models, cross entropy is used as a loss function in the training process, and training the first detection model and the second detection model is completed after the loss function reaches a preset convergence condition.
It should be noted that, compared with the second detection model, for the image without texture information, the first detection model can complete the appearance defect detection of the related image at a faster speed, and the defect can be determined to exist only by the abnormal texture information in the image, so that the detection efficiency is higher. And for the image with the texture information, in the process of training the second detection model, longer time is needed, because the texture information can cause interference to the detection process, the first detection model and the second detection model obtained by training in the mode have higher detection speed and higher detection precision, and the appearance detection of the mobile phone shell is completed.
Referring to fig. 2, a schematic structural diagram of an artificial intelligence based appearance detection system for a mobile phone case according to an embodiment of the present invention is provided, as another aspect of the present invention, an artificial intelligence based appearance detection system for a mobile phone case is provided, which includes:
the mobile phone shell scanning module scans the mobile phone shell to be detected to generate a shell image to be detected;
specifically, the mobile phone shell scanning module can scan the mobile phone shell to be detected under the irradiation of the LED light source through a CMOS camera or a CCD camera to obtain an image of the shell to be detected, and in the embodiment, the CCD camera is taken as an example;
the image preprocessing module is used for carrying out gray processing on the shell image to be detected to obtain a characteristic gray image;
the image segmentation module is used for dividing the characteristic gray image into a plurality of image areas;
the feature analysis module is used for calculating the texture feature value of each image area of the feature gray level image;
specifically, the calculation of the texture feature value includes the following steps:
for any image area, processing the image area based on a circular LBP operator to obtain an LBP characteristic image, taking an LBP value of each pixel point in the LBP characteristic image as a gray value, determining a gray level co-occurrence matrix of the LBP characteristic image, extracting texture features of the image area through the gray level co-occurrence matrix, carrying out statistical analysis on the gray level co-occurrence matrix, calculating to obtain entropy of the gray level co-occurrence matrix, and taking the entropy of the gray level co-occurrence matrix as the texture feature value of the image area;
the region screening module is used for screening a plurality of image regions of the characteristic gray level image;
the appearance detection module is used for detecting appearance defects of the image areas obtained by the area screening module to obtain detection results of the mobile phone shell to be detected;
specifically, the appearance detection module detects appearance defects of the mobile phone shell to be detected through the first detection model and the second detection model.
It will be understood that modifications and variations will be apparent to those skilled in the art from the foregoing description, and it is intended that all such modifications and variations be included within the scope of the following claims. Parts of the specification not described in detail belong to the prior art known to those skilled in the art.
Claims (7)
1. The method for detecting the appearance of the mobile phone shell based on the artificial intelligence is characterized by comprising the following steps of:
scanning a mobile phone shell to be detected to generate a shell image to be detected, and carrying out graying treatment on the shell image to be detected to obtain a characteristic gray image;
obtaining a standard gray level image of a sample mobile phone shell, wherein the sample mobile phone shell does not have appearance defects, and carrying out region segmentation on the characteristic gray level image based on a preset segmentation size to obtain a plurality of image regions, so as to construct an image region set Z;
traversing each image area in the image area set Z, calculating a texture characteristic value of each image area, and for any image area, marking the image area as a first type area if the texture characteristic value is larger than a preset classification threshold value, otherwise marking the image area as a second type area, and constructing to obtain a first type area set C1 and a second type area set C2;
dividing the characteristic gray image into a plurality of image areas based on the preset dividing size, mapping each image area in the first type area set C1 and the second type area set C2 into the characteristic gray image, and determining the area type of each image area in the characteristic gray image;
and for the characteristic gray level image, sending the image area belonging to the first type area to a trained first detection model for defect detection, calculating the texture characteristic value of each image area for the image area belonging to the second type area, respectively calculating the texture difference value of the image area belonging to the second type area in the characteristic gray level image and the image area at the same position in the standard gray level image through the texture characteristic value, and sending the image area with the texture difference value larger than a preset difference threshold value in the characteristic gray level image to a second detection model for defect detection to obtain the detection result of the mobile phone shell to be detected.
2. The method of claim 1, wherein for the feature gray scale image and the standard gray scale image, further comprising:
the characteristic gray level image and the standard gray level image are images after filling the hole area and the logo area through mask operation;
scanning the sample mobile phone shell and performing image graying treatment to obtain a sample shell gray image, performing background segmentation treatment and connected domain analysis on the sample shell gray image, determining a plurality of connected domains of the sample shell gray image, screening the connected domains according to a preset size threshold, determining a hole area and a logo area of the sample mobile phone shell, performing filling treatment on the hole area and the logo area on the sample shell gray image through mask operation to obtain the standard gray image, and performing graying treatment on the shell image to be detected to obtain the shell gray image to be detected;
mapping the hole area and logo area of the sample mobile phone shell into the gray level image of the shell to be detected, and replacing gray level values of the hole area and logo area of the gray level image of the shell to be detected with filling content of the hole area and logo area of the sample mobile phone shell to obtain the characteristic gray level image.
3. The method according to claim 1, wherein calculating the texture feature value of each image region comprises:
for any image area, processing the image area based on a circular LBP operator to obtain an LBP characteristic image, taking an LBP value of each pixel point in the LBP characteristic image as a gray level value, determining a gray level co-occurrence matrix of the LBP characteristic image, extracting texture features of the image area through the gray level co-occurrence matrix, carrying out statistical analysis on the gray level co-occurrence matrix, calculating to obtain entropy of the gray level co-occurrence matrix, and taking the entropy of the gray level co-occurrence matrix as the texture feature value of the image area.
4. The method of claim 1, comprising the training steps for the first detection model and the second detection model of:
the method comprises the steps of obtaining a training sample image set, wherein the training sample image set comprises a plurality of images which have appearance defects and belong to a first type region and a plurality of images which have appearance defects and belong to a second type region, respectively carrying out defect marking on each image in the training sample image set, classifying the marked training sample images according to the type of the region, constructing a first type training data set and a second type training data set, inputting the first type training data set into a first detection model which is constructed in advance, inputting the second type training data set into a second detection model which is constructed in advance, using cross entropy as a loss function in the training process, and completing training on the first detection model and the second detection model after the loss function reaches a preset convergence condition.
5. The method of claim 1, further comprising, for the housing image to be detected:
and after the shell image to be detected is subjected to gray processing, carrying out noise reduction processing on the shell image to be detected after gray processing through a Gaussian filter algorithm.
6. An artificial intelligence based appearance detection system for a mobile phone casing, wherein the system is configured to implement an artificial intelligence based appearance detection method according to any one of claims 1 to 5, and the system comprises:
the mobile phone shell scanning module scans the mobile phone shell to be detected to generate a shell image to be detected;
the image preprocessing module is used for carrying out gray processing on the shell image to be detected to obtain a characteristic gray image;
the image segmentation module is used for dividing the characteristic gray image into a plurality of image areas;
the feature analysis module is used for calculating the texture feature value of each image area of the feature gray level image;
the region screening module is used for screening a plurality of image regions of the characteristic gray level image;
and the appearance detection module is used for detecting appearance defects of the image areas obtained by the area screening module to obtain detection results of the mobile phone shell to be detected.
7. The system of claim 6, wherein for the appearance detection module, further comprising:
and the appearance detection module detects appearance defects of the mobile phone shell to be detected through the first detection model and the second detection model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311496504.2A CN117237340B (en) | 2023-11-10 | 2023-11-10 | Method and system for detecting appearance of mobile phone shell based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311496504.2A CN117237340B (en) | 2023-11-10 | 2023-11-10 | Method and system for detecting appearance of mobile phone shell based on artificial intelligence |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117237340A CN117237340A (en) | 2023-12-15 |
| CN117237340B true CN117237340B (en) | 2024-01-26 |
Family
ID=89098579
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311496504.2A Active CN117237340B (en) | 2023-11-10 | 2023-11-10 | Method and system for detecting appearance of mobile phone shell based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117237340B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106372645A (en) * | 2016-08-29 | 2017-02-01 | 广东工业大学 | Mobile phone housing complex texture background defect detection method |
| CN113160192A (en) * | 2021-04-28 | 2021-07-23 | 北京科技大学 | Visual sense-based snow pressing vehicle appearance defect detection method and device under complex background |
| CN113588656A (en) * | 2021-07-02 | 2021-11-02 | 江苏恒裕泰智能装备有限公司 | Novel bolt appearance size detection method |
| CN114419004A (en) * | 2022-01-21 | 2022-04-29 | 佛山技研智联科技有限公司 | Fabric flaw detection method and device, computer equipment and readable storage medium |
| CN114494179A (en) * | 2022-01-24 | 2022-05-13 | 深圳闪回科技有限公司 | Mobile phone back damage point detection method and system based on image recognition |
| CN115249246A (en) * | 2022-09-23 | 2022-10-28 | 深圳市欣冠精密技术有限公司 | Optical glass surface defect detection method |
| CN116805416A (en) * | 2023-08-21 | 2023-09-26 | 中国电建集团华东勘测设计研究院有限公司 | Drainage pipeline defect identification model training method and drainage pipeline defect identification method |
| CN116934746A (en) * | 2023-09-14 | 2023-10-24 | 常州微亿智造科技有限公司 | Scratch defect detection method, system, equipment and medium thereof |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109829906A (en) * | 2019-01-31 | 2019-05-31 | 桂林电子科技大学 | It is a kind of based on the workpiece, defect of the field of direction and textural characteristics detection and classification method |
| GB2597857B (en) * | 2019-03-15 | 2023-05-17 | Invista Textiles Uk Ltd | Yarn quality control |
| CN111415343B (en) * | 2020-03-18 | 2020-10-27 | 珠海市奥德维科技有限公司 | Artificial intelligence-based six-side appearance detection method for chip multilayer ceramic capacitor |
| CN112907528B (en) * | 2021-02-09 | 2021-11-09 | 南京航空航天大学 | Point cloud-to-image-based composite material laying wire surface defect detection and identification method |
-
2023
- 2023-11-10 CN CN202311496504.2A patent/CN117237340B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106372645A (en) * | 2016-08-29 | 2017-02-01 | 广东工业大学 | Mobile phone housing complex texture background defect detection method |
| CN113160192A (en) * | 2021-04-28 | 2021-07-23 | 北京科技大学 | Visual sense-based snow pressing vehicle appearance defect detection method and device under complex background |
| CN113588656A (en) * | 2021-07-02 | 2021-11-02 | 江苏恒裕泰智能装备有限公司 | Novel bolt appearance size detection method |
| CN114419004A (en) * | 2022-01-21 | 2022-04-29 | 佛山技研智联科技有限公司 | Fabric flaw detection method and device, computer equipment and readable storage medium |
| CN114494179A (en) * | 2022-01-24 | 2022-05-13 | 深圳闪回科技有限公司 | Mobile phone back damage point detection method and system based on image recognition |
| CN115249246A (en) * | 2022-09-23 | 2022-10-28 | 深圳市欣冠精密技术有限公司 | Optical glass surface defect detection method |
| CN116805416A (en) * | 2023-08-21 | 2023-09-26 | 中国电建集团华东勘测设计研究院有限公司 | Drainage pipeline defect identification model training method and drainage pipeline defect identification method |
| CN116934746A (en) * | 2023-09-14 | 2023-10-24 | 常州微亿智造科技有限公司 | Scratch defect detection method, system, equipment and medium thereof |
Non-Patent Citations (4)
| Title |
|---|
| A Method for Identifying Flaws on Product Surface Based on Spatial Connectivity Domain;Quanyou Zhang etc.;《IEEE Access 》;第2169-3536页 * |
| Effective texture-defect-detecting method;Wen Zhao-lin etc.;《Application Research of Computers》;第57-59页 * |
| 基于机器视觉的钢轨表面缺陷检测系统研究;李宁;《中国优秀硕士学位论文全文数据库（工程科技Ⅱ辑）》;第C033-54页 * |
| 面向织物缺陷检测的CUDA并行图像处理模型与算法研究;司小书;《中国博士学位论文全文数据库（信息科技辑）》;第I138-63页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117237340A (en) | 2023-12-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110148130B (en) | Method and device for detecting part defects | |
| CN108562589A (en) | A method of magnetic circuit material surface defect is detected | |
| CN112862770B (en) | Defect analysis and diagnosis system, method and device based on artificial intelligence | |
| CN110942013A (en) | Satellite image feature extraction method and system based on deep neural network | |
| CN112614062A (en) | Bacterial colony counting method and device and computer storage medium | |
| CN105701491A (en) | Method for making fixed-format document image template and application thereof | |
| CN116503388B (en) | Defect detection method, device and storage medium | |
| CN117094975A (en) | Method and device for detecting surface defects of steel and electronic equipment | |
| CN115294377A (en) | System and method for identifying road cracks | |
| CN113392819B (en) | Batch academic image automatic segmentation and labeling device and method | |
| CN113129265B (en) | Method and device for detecting surface defects of ceramic tiles and storage medium | |
| CN117237340B (en) | Method and system for detecting appearance of mobile phone shell based on artificial intelligence | |
| CN116543414A (en) | Tongue color classification and tongue redness and purple quantification method based on multi-model fusion | |
| CN112750113B (en) | Glass bottle defect detection method and device based on deep learning and linear detection | |
| JP3906221B2 (en) | Image processing method and image processing apparatus | |
| CN111681185B (en) | Finite element modeling method based on X-ray scanning image of asphalt mixture | |
| CN116563237B (en) | Deep learning-based chicken carcass defect hyperspectral image detection method | |
| CN116958126B (en) | Method and system for detecting injection molding of mold based on computer vision | |
| CN115222732B (en) | Injection molding process anomaly detection method based on big data analysis and color difference detection | |
| CN109087289A (en) | A kind of plate visual pattern detection algorithm under grid background | |
| TWI781835B (en) | Fabric information digitization system and method thereof | |
| CN116311262B (en) | Instrument information identification method, system, equipment and storage medium | |
| JP6961319B2 (en) | Evaluation method and evaluation device | |
| CN117058089A (en) | Cigarette appearance detection method | |
| Cui | Research on Identification System of Aero Engine Blade |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |