CN107203990B - Label breakage detection method based on template matching and image quality evaluation - Google Patents
Label breakage detection method based on template matching and image quality evaluation Download PDFInfo
- Publication number
- CN107203990B CN107203990B CN201710215105.2A CN201710215105A CN107203990B CN 107203990 B CN107203990 B CN 107203990B CN 201710215105 A CN201710215105 A CN 201710215105A CN 107203990 B CN107203990 B CN 107203990B
- Authority
- CN
- China
- Prior art keywords
- image
- label
- template
- detected
- area
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
- G06T7/001—Industrial image inspection using an image reference approach
Abstract
The invention discloses a label breakage detection method based on template matching and image quality evaluation. For an image to be detected and a template image thereof, a binary image and a gray image of a labeling area are accurately extracted through a series of preprocessing means, shape distortion possibly caused by angle deviation is corrected through affine transformation, position matching of a lossless image and the image to be detected is realized through template matching, a possible defect area is finally obtained through a differential image of the binary image, a real defect area is effectively detected through an image quality evaluation method combining structural similarity and a perception hash value, and accurate positioning of a damaged defect is realized. The method has the innovation points that a simple and efficient preprocessing means is used for extracting the ROI area, the image quality evaluation algorithm is used for calculating the image similarity, and the accuracy and the efficiency of detection are guaranteed.
Description
Technical Field
The invention relates to the technical field of machine vision and video image processing, in particular to a label breakage detection method based on template matching and image quality evaluation.
Background
The traditional bottle label defect detection is mainly used for completing the rear-end detection work of a filling product production line by a manual detection method, and has the problems of low efficiency, high cost, poor stability and reliability and the like, which is greatly inconsistent with large-scale industrial production. The image detection algorithm based on machine vision has a good development prospect, can automatically detect the labeling defects, and effectively solves the problem.
In the current research situation at home and abroad, algorithms commonly used for labeling defects include an image filtering-based method, an edge extraction-based method, a deep learning-based method and the like. However, these methods usually have strict conditions for image capturing, complicated operating environment, and also are not ideal in processing speed, and most of them have certain limitations in practical industrial applications.
The invention provides a label damage detection method based on template matching and image quality evaluation. The method comprises the steps of accurately extracting a binary image and a gray image of a labeling area through a series of preprocessing means, correcting shape distortion possibly caused by angle deviation by affine transformation, realizing position matching of a lossless image and an image to be detected by template matching, finally obtaining a possible defect area through a differential image of the binary image, effectively detecting a real defect area by an image quality evaluation method combining structural similarity and a perception hash value, and accurately positioning the damaged defect.
Disclosure of Invention
The invention aims to solve the problems that: the existing bottle label damage defect detection system relies on manual detection, and has low efficiency and poor stability; the existing various label damage defect detection methods based on machine vision often have strict conditions for image shooting and complex operation environment. The invention provides a label damage defect method based on template matching and image quality evaluation, which not only allows the deviation of a bottle body placing angle within a certain range, but also has high processing speed, and can accurately detect the tiny tilting defects.
The technical scheme of the invention is as follows: a label damage detection method based on template matching and image quality evaluation is used for detecting labels on the surfaces of bottles, a nondestructive image of a label is used as a template, after a label image to be detected is collected, a binary image and a gray image of a label area are extracted from the label image and the label image, shape distortion caused by shooting angle deviation is corrected by affine transformation, position matching of the nondestructive image and the image to be detected is achieved by template matching, a suspected defect area is obtained through a differential image of the binary image, on the basis, a defect area is detected and determined by an image quality evaluation method combining structural similarity and a perception hash value, and accurate positioning of damage defects is achieved.
The invention comprises the following steps:
step one, image acquisition, namely acquiring bottle label images by utilizing an infrared light source, wherein the acquired original infrared images are gray images, and the acquired lossless template images are recorded as Fm(x, y) the label image to be detected is Fo(x, y), wherein subscript m represents a non-destructive label template image and subscript o represents a label image to be detected;
step two, the initial image F obtained in the step onem(x, y) and Fo(x, y) carrying out binarization processing, and removing light spots of non-labeling areas through morphological corrosion operation to obtain a nondestructive labeling binarization image Bm(x, y) and the label binary image B to be detectedo(x,y)；
Step three, utilizing the binary image B obtained in the step twom(x, y) and Bo(x, y) obtaining a binary image RB of the rectangular area where the label is located by taking the minimum external rectangle of the (x, y)m(x, y) and RBo(x, y), and extracting a rectangular area image of the corresponding label in the initial gray level image by using the position information, and recording the rectangular area image as Rm(x, y) and Ro(x,y)；
Step four, obtaining a binaryzation image RB of the to-be-detected labeling area in the step threeo(x, y) and grayscale images Ro(x, y) correcting the shape, and matching the corrected image with the corresponding non-destructive labeling moldTemplate matching is carried out on the plate image, the template matching adopts a normalized square difference matching mode to obtain a label rectangular region gray image and a label rectangular region binary image which are consistent in size, and the gray image is marked as M after the size of the template image is processedm(x, y), the binarized image is denoted as MBm(x, y), the gray scale image after the image to be measured is processed is marked as Mo(x, y), the binarized image is denoted as MBo(x,y)；
Step five, the corrected label rectangular area binaryzation image MB obtained in the step fourm(x, y) and MBo(x, y) carrying out differential operation, removing interference by using morphological processing to obtain a differential image D (x, y), wherein the differential image D (x, y) comprises a plurality of communication areas, taking a minimum external rectangle for each communication area, and labeling a rectangular area gray level image M after correctionm(x, y) and Mo(x, y) extracting corresponding gray scale images Smi(x, y) and Soi(x, y), performing similarity calculation based on image quality evaluation, and judging the communication area as a defect area if a set threshold value is exceeded, wherein the subscript i represents the serial number of the currently processed communication area.
The first step is specifically as follows: collecting images by using a CCD camera in a low-brightness environment, wherein the light source is an infrared light source and is respectively polished from the left side and the right side of the bottle body, and the included angle between the polishing direction and the plane where the bottle body label is located is 45 degrees; the central axis direction of the camera is in the same direction with the normal direction of the plane where the bottle body label is positioned, the deviation is not more than 10 degrees, the label area is positioned in the central area of the shooting range of the camera, and a non-damaged template picture F is collectedm(x, y) and to-be-inspected image Fo(x,y)。
In the fourth step, the binaryzation image RB of the label area to be detected obtained in the third stepo(x, y) and grayscale images Ro(x, y) performing shape correction based on the template image using affine transformation, respectively: extracting characteristic points of the template picture and the picture to be detected by using FAST corner detection, performing binary description by using a BRIEF algorithm, then matching the characteristic points, eliminating mismatching points, calculating an affine transformation matrix between the description template picture and the picture to be detected, and calculating a corrected image by using the matrix.
And calculating the similarity based on the image quality evaluation in the step five as follows: calculating the corresponding template graph and the graph to be detected by using a mode of combining the structural similarity SSIM and the perception hash value, judging that the similarity of the two graphs is smaller when one of the two indexes exceeds a threshold value, namely the existence of defects, and specifically judging as follows:
x, Y, respectively representing the image to be detected and the corresponding template image, result represents the final judgment result, 1 represents that the current image has defects, and 0 represents that the current image has no defects; s is a structural similarity threshold, and p is a perceptual hash threshold.
The invention provides a label damage detection method based on template matching and image quality evaluation, which not only meets the requirement of real-time performance, but also has higher accuracy, and allows the deviation of the bottle body placement angle within a certain range. The innovation points are as follows: the method extracts the label ROI area by using a simple and efficient preprocessing means, and then calculates the similarity between the image to be detected and the corresponding template image by using an image quality evaluation algorithm so as to judge whether the defect exists. Therefore, the accuracy of detection can be ensured, and the detection efficiency can be greatly improved.
Drawings
FIG. 1 is a block flow diagram of the method of the present invention.
Fig. 2 and fig. 3 are the detection results of the damaged label in the method of the present invention, wherein fig. 2 shows the defect of label damage, and fig. 3 shows the defect of label lift. In the two sets of images, (a1) (b1) is a lossless template image, (a2) (b2) is a lossy label original image, (a3) (b3) is an original image of a rectangular region where the corresponding label is located, and (a4) (b4) is a detection result image.
Detailed Description
For a further understanding of the present invention, reference will now be made in detail to the following examples and accompanying drawings.
The invention provides a novel method for detecting the damage defect of a label on the surface of a bottle body, which can quickly and effectively realize the automatic detection of the damage defect of the label on the bottle body. The method mainly comprises four parts of image acquisition, image preprocessing, ROI area extraction and image quality evaluation.
As shown in FIG. 1, the label damage detection method based on template matching and image quality evaluation provided by the invention accurately extracts a binary image and a gray image of a label area through a series of preprocessing means, corrects possible shape distortion caused by angle deviation by affine transformation, realizes position matching of a lossless image and an image to be detected by template matching, finally obtains a possible defect area through a differential image of the binary image, and effectively detects a real defect area by using an image quality evaluation method combining structural similarity and a sensed hash value, thereby realizing accurate positioning of damage defects.
The following describes the specific implementation method of this embodiment in detail:
1. image acquisition:
collecting bottle label image by infrared light source in low brightness environment, wherein the collected original infrared image is gray scale image and the non-destructive template image is Fm(x, y) inspection map Fo(x, y), wherein the subscript m represents the non-destructive label template image and the subscript o represents the label image to be detected.
Specifically, a CCD camera is used for collecting images, an infrared light source is selected as the light source, the light source is respectively polished from the left side and the right side of a bottle body label, the included angle between the light source and the plane where the bottle body label is located is 45 degrees, the invention aims at the plane label, as shown in figures 2 and 3, the plane label can have a slight radian but can be ignored, and the image is processed according to the plane when being processed; the central axis direction of the camera is in the same direction with the normal direction of the plane where the bottle body label is located, certain angle deviation is tolerated in practical application, but the included angle between the normal direction of the plane where the label to be detected is required to be ensured to be not more than 10 degrees and the central axis of the camera, the label area is located in the central area of the shooting range of the camera, and a non-damaged template picture F is collectedm(x, y) and to-be-inspected image Fo(x,y)。
2. Image preprocessing:
the image preprocessing comprises a plurality of steps, and finally label images of the image to be detected and the template image, which correspond to each other in position and are consistent in size, are obtained through a series of operations, wherein the label images comprise a gray level image and a binary image. The method comprises the following specific steps:
(1) for the initial gray image Fm(x, y) and Fo(x, y) carrying out binarization processing, and removing some slender strip-shaped light spot interference possibly appearing in the non-labeling area on the surface of the bottle body through morphological corrosion operation, wherein the light spots are usually caused by bottle body shape bulges, thereby obtaining a nondestructive labeling binarization image Bm(x, y) and label binary image B to be detectedo(x,y)；
(2) Taking a binary image Bm(x, y) and Bo(x, y) minimum bounding rectangle to obtain the binary image RB of the rectangular region where the label is positionedm(x, y) and RBo(x, y), and using the position information to extract the rectangular area image of the corresponding label in the initial gray level image, and recording as Rm(x, y) and Ro(x,y)；
(3) Binaryzation image RB of label area to be detectedo(x, y) and grayscale images Ro(x, y) carrying out shape correction, specifically, extracting characteristic points of the template picture and the picture to be detected by using FAST corner detection, carrying out binary description by using a BRIEF algorithm, then matching the characteristic points, eliminating mismatching points, calculating an affine transformation matrix between the template picture and the picture to be detected, and calculating a corrected image by using the matrix.
After image correction, respectively carrying out template matching on the corrected image of the image to be detected and the corresponding template image, wherein the template matching adopts a normalized square error matching mode to obtain a label rectangular area gray level image and a label rectangular area binary image which are consistent in size, and the gray level image after template image processing is marked as Mm(x, y), the binarized image is denoted as MBm(x, y), and marking the gray level image after the image to be detected is processed as Mo(x, y), the binarized image is denoted as MBo(x, y). The template image is processed because the two images need to be the same in size, and the situation that the label to be detected is larger than the template label in practical application may occur, such as the situation that the label is tilted, at the moment, the template image needs to be increased in size, and the value of the pixel point in the region is increased to be0。
3. Extracting an ROI (region of interest):
the ROI region is a region where a breakage defect may exist. The obtained corrected label rectangular area binaryzation figure MBm(x, y) and MBoAnd (x, y) carrying out difference operation, and carrying out morphological corrosion treatment on the obtained difference image to remove interference to obtain a difference image D (x, y). The difference image D (x, y) may include a plurality of connected regions, and a minimum bounding rectangle is taken for each connected region, which is the ROI region. Labeling the rectangular region gray level image M after correctionm(x, y) and Mo(x, y) extracting corresponding gray scale images Smi(x, y) and Soi(x, y), subscript i indicates the serial number of the currently processed connected region.
4. And (3) image quality evaluation:
for gray level image Smi(x, y) and Soi(x, y) carrying out image quality evaluation, and judging the communication area as a defect area by using a method for calculating similarity and exceeding a threshold value.
The image similarity calculation specific algorithm is as follows: and calculating the corresponding template graph and the graph to be detected by utilizing a mode of combining the structural similarity SSIM and the perception hash value, and judging that the similarity of the two graphs is smaller if one of the two indexes exceeds a threshold value, namely the defect exists. The specific threshold setting is different due to different detection objects, and can be set through positive and negative samples according to actual needs. The specific determination method is as follows:
x, Y respectively represent the images to be detected and their corresponding template images. Wherein result represents the final judgment result, 1 represents that the current image has defects, and 0 represents that the image has no defects; s is a structural similarity threshold, and p is a perceptual hash threshold.
The similarity calculation method is specifically as follows.
(1) The structural similarity SSIM measures the image similarity from three aspects of brightness, contrast and structure, and the specific calculation method is as follows:
SSIM(X,Y)＝l(X,Y)*c(X,Y)*s(X,Y) (2)
wherein:
x, Y denote two images for which the similarity is to be calculated, where μX、μYRespectively representing the mean, σ, of the image X, YX、 σYRespectively representing the variance, σ, of the image X, YXYRepresents the covariance of images X and Y; c1、C2、C3Is a constant, take C1＝(K1*L)2，C2＝(K2*L)2，C3＝σXY 2/2, taking K1＝0.01，K2＝0.03，L＝255。
The SSIM value range is [0,1], and the smaller the value, the lower the similarity is, namely the higher the possibility that the image has defects.
(2) The perception hash value measures the image similarity from the sense of human eyes. In order to remove redundant image information, improve detection efficiency and effectively avoid interference caused by histogram equalization and the like which may exist, a DCT-based hash calculation method is specifically adopted, and the specific calculation method is as follows:
the fingerprints of the two images are respectively calculated: reducing the image size to 32 x 32; calculating DCT transformation of the picture to obtain a 32 x 32 DCT coefficient matrix; as long as the matrix of 8 × 8 at the upper left corner is reserved, the part is image low-frequency information; calculating a mean value of the numerical values in the matrix; 8, setting the DCT matrix to be 1 when the DCT matrix is higher than the average value and setting the DCT matrix to be 0 when the DCT matrix is lower than the average value, and obtaining the fingerprint of the image;
and calculating the Hamming distance of the two image fingerprints, namely the hash value. The larger the hash value is, the larger the difference between the two images is, the lower the similarity is, i.e. the higher the possibility of the image having defects is.
Fig. 2 and fig. 3 are diagrams illustrating the effect of the present invention, wherein fig. 2 illustrates the defect of label breakage, and fig. 3 illustrates the defect of label lifting. In the two sets of images, (a1) (b1) is a lossless template image, (a2) (b2) is a lossy label original image, (a3) (b3) is an original image of a rectangular region where the corresponding label is located, and (a4) (b4) is a detection result image. As can be seen from the figure, the method for detecting the label breakage defect provided by the invention can effectively detect the image breakage defect and the tilting defect with a smaller angle, and has a good effect.
The present invention and its embodiments have been described above schematically, without limitation, and what is shown in the drawings is only one of the embodiments of the present invention, and the actual structure is not limited thereto. Therefore, if the person skilled in the art receives the teaching, without departing from the spirit of the invention, the person skilled in the art shall not inventively design the similar structural modes and embodiments to the technical solution, but shall fall within the scope of the invention.
Claims (4)
1. A label damage detection method based on template matching and image quality assessment is used for detecting labels on the surfaces of bottles and is characterized in that a nondestructive image of a label is taken as a template, after a label image to be detected is collected, a binary image and a gray image of a label area are extracted from the label image and the label image, affine transformation is used for correcting shape distortion caused by shooting angle deviation, the template matching is used for realizing position matching of the nondestructive image and the image to be detected, a suspected defect area is obtained through a differential image of the binary image, on the basis, a defect area is detected and determined by using an image quality assessment method combining structural similarity and a perception hash value, and accurate positioning of damage defects is realized;
the method comprises the following steps:
step one, image acquisition, namely acquiring bottle body label images by utilizing an infrared light source, wherein the acquired original infrared images are gray level images, and the acquired lossless template images are recorded as Fm(x, y) the label image to be detected is Fo(x, y), wherein subscript m represents a non-destructive label template image and subscript o represents a label image to be detected;
step two, the initial image F obtained in the step onem(x, y) and Fo(x, y) carrying out binarization processing, and removing light spots of non-labeling areas through morphological corrosion operation to obtain a nondestructive labeling binarization image Bm(x, y) and the label binary image B to be detectedo(x,y)；
Step three, utilizing the binary image B obtained in the step twom(x, y) and Bo(x, y) obtaining a binary image RB of the rectangular area where the label is located by taking the minimum external rectangle of the (x, y)m(x, y) and RBo(x, y), and extracting a rectangular area image of the corresponding label in the initial gray level image by using the position information, and recording the rectangular area image as Rm(x, y) and Ro(x,y)；
Step four, obtaining a binaryzation image RB of the to-be-detected labeling area in the step threeo(x, y) and grayscale images Ro(x, y) carrying out shape correction, carrying out template matching on the corrected image and a corresponding lossless label template image, wherein the template matching adopts a normalized square error matching mode to obtain a label rectangular area gray image and a label rectangular area binary image which are consistent in size, and the gray image after the size processing of the template image is marked as Mm(x, y), the binarized image is denoted as MBm(x, y), and marking the processed gray image of the image to be detected as Mo(x, y), the binarized image is denoted as MBo(x,y)；
Step five, the corrected label rectangular area binaryzation image MB obtained in the step fourm(x, y) and MBo(x, y) carrying out difference operation, removing interference by using morphological processing to obtain a difference image D (x, y), wherein the difference image D (x, y) comprises a plurality of communication areas, a minimum external rectangle is taken for each communication area, and a rectangular area gray level image M is labeled after correctionm(x, y) and Mo(x, y) extracting corresponding gray scale images Smi(x, y) and Soi(x, y) performing similarity calculation based on image quality evaluation, and determining the link region as a defect region if a set threshold value is exceeded, wherein the subscript i represents the link region currently processedA serial number.
2. The method for detecting label damage based on template matching and image quality evaluation as claimed in claim 1, wherein the first step is specifically as follows: collecting images by using a CCD camera in a low-brightness environment, wherein the light source is an infrared light source and is respectively polished from the left side and the right side of the bottle body label, and the included angle between the polishing direction and the plane where the bottle body label is located is 45 degrees; the central axis direction of the camera is in the same direction with the normal direction of the plane where the bottle body label is positioned, the deviation is not more than 10 degrees, the label area is positioned in the central area of the shooting range of the camera, and a non-damaged template picture F is collectedm(x, y) and to-be-inspected image Fo(x,y)。
3. The method for detecting label damage based on template matching and image quality evaluation as claimed in claim 1, wherein in the fourth step, the binarized image RB of the label region to be detected obtained in the third stepo(x, y) and grayscale images Ro(x, y) performing shape correction based on the template image using affine transformation, respectively: extracting characteristic points of the template picture and the picture to be detected by using FAST corner detection, performing binary description by using a BRIEF algorithm, then matching the characteristic points, eliminating mismatching points, calculating an affine transformation matrix between the description template picture and the picture to be detected, and calculating a corrected image by using the matrix.
4. The label breakage detection method based on template matching and image quality evaluation as claimed in claim 1, wherein the similarity based on image quality evaluation in the fifth step is calculated as: the method comprises the following steps of calculating a corresponding template graph and a graph to be detected by utilizing a mode of combining structural similarity SSIM and a perception hash value, judging whether defects exist when one of two indexes exceeds a threshold value, and specifically judging as follows:
x, Y, respectively representing the image to be detected and the template image corresponding to the image to be detected, result represents the final judgment result, 1 represents that the current image has defects, and 0 represents that the current image has no defects; s is a structural similarity threshold, and p is a perceptual hash threshold.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710215105.2A CN107203990B (en) | 2017-04-02 | 2017-04-02 | Label breakage detection method based on template matching and image quality evaluation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710215105.2A CN107203990B (en) | 2017-04-02 | 2017-04-02 | Label breakage detection method based on template matching and image quality evaluation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107203990A CN107203990A (en) | 2017-09-26 |
| CN107203990B true CN107203990B (en) | 2020-12-18 |
Family
ID=59906453
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710215105.2A Active CN107203990B (en) | 2017-04-02 | 2017-04-02 | Label breakage detection method based on template matching and image quality evaluation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107203990B (en) |
Families Citing this family (24)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107767379B (en) * | 2017-11-16 | 2021-02-12 | 桂林电子科技大学 | PCB label printing quality detection method |
| CN108230324B (en) * | 2018-01-31 | 2023-10-20 | 浙江理工大学 | Visual detection method for microdefect on surface of magnetic shoe |
| CN108229583B (en) * | 2018-02-02 | 2022-04-29 | 北京邮电大学 | Method and device for fast template matching based on main direction difference characteristics |
| CN108327398A (en) * | 2018-03-30 | 2018-07-27 | 湖北工程学院 | It is merged facial mask detection method and device |
| CN108596905A (en) * | 2018-05-10 | 2018-09-28 | 广东拓斯达科技股份有限公司 | Object detecting method, system, computer equipment and storage medium |
| CN108871185B (en) * | 2018-05-10 | 2020-12-29 | 苏州大学 | Method, device and equipment for detecting parts and computer readable storage medium |
| CN108961252B (en) * | 2018-07-27 | 2021-06-08 | Oppo（重庆）智能科技有限公司 | Method for detecting logo paste failure, electronic device and computer readable storage medium |
| CN109670070B (en) * | 2018-11-29 | 2022-07-29 | 昆明理工大学 | Image similarity judgment method based on double Hash |
| CN109658584B (en) * | 2018-12-14 | 2021-01-12 | 泰康保险集团股份有限公司 | Bill information identification method and device |
| CN109636797B (en) * | 2018-12-21 | 2022-08-05 | 国网福建省电力有限公司 | Method for detecting deformation of non-contact flexible direct current converter valve module |
| CN109801287B (en) * | 2019-01-30 | 2023-04-28 | 温州大学 | Label breakage detection method based on template matching and image quality evaluation |
| CN109949479A (en) * | 2019-03-18 | 2019-06-28 | 成都好享你网络科技有限公司 | Data processing method, device and intelligent vending machine based on image difference |
| CN110047063B (en) * | 2019-03-21 | 2023-05-16 | 深圳市燕麦科技股份有限公司 | Material drop detection method, device, equipment and storage medium |
| CN109894387A (en) * | 2019-03-29 | 2019-06-18 | 江苏省烟草公司无锡市公司 | A kind of cigarette stamp detection method, apparatus and system |
| CN110246122A (en) * | 2019-05-20 | 2019-09-17 | 江苏理工学院 | Small size bearing quality determining method, apparatus and system based on machine vision |
| CN112308814A (en) * | 2019-07-26 | 2021-02-02 | 北京四方继保自动化股份有限公司 | Method and system for automatically identifying switch on-off position state of disconnecting link of power system |
| CN113469921B (en) * | 2021-09-06 | 2022-02-08 | 深圳市创世易明科技有限公司 | Image defect repairing method, system, computer device and storage medium |
| CN114219803B (en) * | 2022-02-21 | 2022-07-15 | 浙江大学 | Detection method and system for three-stage image quality evaluation |
| CN114445472B (en) * | 2022-03-04 | 2023-05-26 | 山东胜算软件科技有限公司 | Multi-step image registration method based on affine transformation and template matching |
| CN114677336B (en) * | 2022-03-10 | 2024-04-09 | 四川省建筑科学研究院有限公司 | Curtain wall panel damage identification method based on infrared image |
| CN114851206B (en) * | 2022-06-06 | 2024-03-29 | 天津中科智能识别有限公司 | Method for grabbing stove based on vision guiding mechanical arm |
| CN115205288B (en) * | 2022-09-14 | 2023-01-03 | 江苏智云天工科技有限公司 | Industrial defect detection method and device |
| CN116681695B (en) * | 2023-07-27 | 2023-12-01 | 山东阁林板建材科技有限公司 | Quality detection method for anti-deformation template end face |
| CN117197128B (en) * | 2023-11-02 | 2024-01-30 | 宁波明禾新能源科技有限公司 | Product image detection method and system of bending machine and bending machine |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103426191A (en) * | 2012-05-26 | 2013-12-04 | 百度在线网络技术（北京）有限公司 | Method and system for picture marking |
| CN104318259A (en) * | 2014-10-20 | 2015-01-28 | 北京齐尔布莱特科技有限公司 | Target picture identifying device and method for and computing device |
| CN105160654A (en) * | 2015-07-09 | 2015-12-16 | 浙江工商大学 | Towel label defect detecting method based on feature point extraction |
| CN105956020A (en) * | 2016-04-22 | 2016-09-21 | 广州精点计算机科技有限公司 | Method and device for searching similar image based on perceptual hash |
| CN106040608A (en) * | 2016-05-20 | 2016-10-26 | 上海交通大学 | Dynamic optical detection device and detection method for laptop keyboards |
-
2017
- 2017-04-02 CN CN201710215105.2A patent/CN107203990B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103426191A (en) * | 2012-05-26 | 2013-12-04 | 百度在线网络技术（北京）有限公司 | Method and system for picture marking |
| CN104318259A (en) * | 2014-10-20 | 2015-01-28 | 北京齐尔布莱特科技有限公司 | Target picture identifying device and method for and computing device |
| CN105160654A (en) * | 2015-07-09 | 2015-12-16 | 浙江工商大学 | Towel label defect detecting method based on feature point extraction |
| CN105956020A (en) * | 2016-04-22 | 2016-09-21 | 广州精点计算机科技有限公司 | Method and device for searching similar image based on perceptual hash |
| CN106040608A (en) * | 2016-05-20 | 2016-10-26 | 上海交通大学 | Dynamic optical detection device and detection method for laptop keyboards |
Non-Patent Citations (2)
| Title |
|---|
| 基于机器视觉的印刷质量在线自动检测系统研究;郭健;《中国优秀硕士学位论文全文数据库工程科技Ⅰ辑》;20170215;正文第4页第2段-9第页,第12-17页,第33-89页 * |
| 基于机器视觉的标签缺陷检测系统设计与实现;汪清芳;《中国优秀硕士学位论文全文数据库信息科技辑》;20160315;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107203990A (en) | 2017-09-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107203990B (en) | Label breakage detection method based on template matching and image quality evaluation | |
| CN110314854B (en) | Workpiece detecting and sorting device and method based on visual robot | |
| CN110220917B (en) | Crown cap surface defect online detection method based on image processing | |
| CN105067638B (en) | Tire fetal membrane face character defect inspection method based on machine vision | |
| CN108776140B (en) | Machine vision-based printed matter flaw detection method and system | |
| CN109785316B (en) | Method for detecting apparent defects of chip | |
| CN104504388B (en) | A kind of pavement crack identification and feature extraction algorithm and system | |
| WO2017012581A1 (en) | Method and system for decoding qr code based on weighted average grey method | |
| CN110033431B (en) | Non-contact detection device and detection method for detecting corrosion area on surface of steel bridge | |
| CN107561087A (en) | A kind of mouse logo positioning and defect inspection method based on machine vision | |
| CN108805862B (en) | Label identification method based on improved structure similarity | |
| CN112037203A (en) | Side surface defect detection method and system based on complex workpiece outer contour registration | |
| US20140301608A1 (en) | Chemical structure recognition tool | |
| CN105403147A (en) | Embedded bottle pre-form detection system and detection method | |
| CN111667475B (en) | Machine vision-based Chinese date grading detection method | |
| CN110648330B (en) | Defect detection method for camera glass | |
| CN113251926B (en) | Method and device for measuring size of irregular object | |
| CN107389693A (en) | A kind of printed matter defect automatic testing method based on machine vision | |
| CN110751619A (en) | Insulator defect detection method | |
| CN106780437B (en) | A kind of quick QFN chip plastic packaging image obtains and amplification method | |
| CN111242888A (en) | Image processing method and system based on machine vision | |
| CN115752969A (en) | Method, system and equipment for detecting sealing performance of aluminum foil seal | |
| CN111968082A (en) | Product packaging defect detection and identification method based on machine vision | |
| CN111638218A (en) | Method for detecting surface defects of coating | |
| CN107507130A (en) | A kind of quickly QFN chip pins image obtains and amplification method |
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 |