CN106204618A - Product surface of package defects detection based on machine vision and sorting technique - Google Patents

Product surface of package defects detection based on machine vision and sorting technique Download PDF

Info

Publication number
CN106204618A
CN106204618A CN201610579970.0A CN201610579970A CN106204618A CN 106204618 A CN106204618 A CN 106204618A CN 201610579970 A CN201610579970 A CN 201610579970A CN 106204618 A CN106204618 A CN 106204618A
Authority
CN
China
Prior art keywords
image
defect
product surface
standard picture
machine vision
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
Application number
CN201610579970.0A
Other languages
Chinese (zh)
Inventor
杨丹婷
欧阳光
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nanjing Wencai Science & Technology Co Ltd
Original Assignee
Nanjing Wencai Science & Technology Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Nanjing Wencai Science & Technology Co Ltd filed Critical Nanjing Wencai Science & Technology Co Ltd
Priority to CN201610579970.0A priority Critical patent/CN106204618A/en
Publication of CN106204618A publication Critical patent/CN106204618A/en
Pending legal-status Critical Current

Links

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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
    • G06F18/24133—Distances to prototypes
    • G06F18/24137—Distances to cluster centroïds
    • G06F18/2414—Smoothing the distance, e.g. radial basis function networks [RBFN]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/42—Global feature extraction by analysis of the whole pattern, e.g. using frequency domain transformations or autocorrelation
    • G06V10/422—Global feature extraction by analysis of the whole pattern, e.g. using frequency domain transformations or autocorrelation for representing the structure of the pattern or shape of an object therefor
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/56—Extraction of image or video features relating to colour
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10024—Color image
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30108—Industrial image inspection
    • G06T2207/30144—Printing quality

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Data Mining & Analysis (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Multimedia (AREA)
  • Quality & Reliability (AREA)
  • Image Processing (AREA)
  • Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
  • Image Analysis (AREA)

Abstract

The present invention discloses a kind of product surface of package defects detection based on machine vision and sorting technique, comprises the steps: that (1) gathers the high clear colorful image of no defective product packaging, makes standard picture；Recycling photographic head captured in real-time, the high clear colorful image of online acquisition product to be measured packaging, as testing image；(2) based on SURF algorithm, testing image and standard picture are carried out images match；(3) two images after coupling in step (2) are differed from shadow operation, obtain defect image；(4) defect image is carried out feature extraction, obtain geometric properties and the color character of defect image；(5) use RBF neural algorithm, product surface of package defect is classified.The present invention carries out automatic defect detection and classification by Vision Builder for Automated Inspection, can avoid interference from human factor, cost of labor be greatly lowered, thus avoids the huge invisible costs such as training, the management that manual detection brings.

Description

Product surface of package defects detection based on machine vision and sorting technique
Technical field
The invention belongs to packaging detection field, be specifically related to the inspection of a kind of product surface of package defect based on machine vision Survey and sorting technique.
Background technology
Along with the development of printing industry in modern age, people are more and more higher to the requirement of printing technology.At the big back of the body of such epoch Under scape, need quick high precision image detection recognizer to meet the high speed of current surface defects detection and high-precision Requirement, from offline inspection to on-line checking, constantly promotes press to develop to high-quality, high efficiency, low cost direction.
Common surface defect mainly has: bites, dirty patch, cut, scumming, fly ink, aberration, register trouble etc..More than Yin Situation causes percent defective too high, can have a strong impact on the quality of production.In order to strictly control disqualification rate, need printing when pair Printing surface is detected, and the most each defective work is rejected.
Traditional detection method identifies mainly by human eye, rejects defective work, but human eye visual inspection easily produces fatigue, The detection slow efficiency of speed is low, and the mark loss that, character little for area is intensive is higher, and cannot ensure unified quality mark Accurate.Therefore, the detection automatically of surface defect is increasingly becoming the trend of industry.
For detecting surface defect, can directly compare the gray value of two width images, such as PatInspect instrument.With treating mapping Image subtraction standard picture, if the difference of corresponding pixel points is more than given threshold value, is labeled as defect.But adopt at actual image During collection, can there is a small amount of deviation in the position of image, and a large amount of defect pixel points that side-play amount causes can cover point, line defect Exist, additionally, the single solution for diverse problems of threshold value affects Detection results.Another detects surface defect system such as DH-CHECK-C printing quality Real-time and On-line, the mass colour that can detect printed patterns is deep or light, lines broken string, word are incomplete, can be to micro text Integrity judges.This system is sampling check, the inspection bitten etc. for Process Character defect, such as register trouble, large area Have certain effect, but this system for fly ink, aberration, cutter silk, bite, the detection of the sudden defect such as white point is not ten Sub-argument is thought.
Additionally, existing most on-line checking is all based on gray level image, and the detection to color is little, but colour cast Problem is a defect problem that can not be ignored.When detecting coloured image, traditional way is that coloured image is turned Turn to gray level image, but this can lose many colouring informations so that the defect relevant with color is difficult to detect.
Therefore, existing surface defects detection system can only detect certain type, a range of defect, has certain Limitation, speed is slow, testing cost is high, detect the shortcomings such as type is single to there is detection.
Summary of the invention
Goal of the invention: present invention aim at for the deficiencies in the prior art, it is provided that a kind of surface defects of products detection with The universal method of classification, Detection accuracy is high, it is adaptable to general machine vision platform.
Technical scheme: product surface of package defects detection based on machine vision of the present invention and sorting technique, it is special Levy and be, comprise the steps:
(1) gather the high clear colorful image of no defective product packaging, make standard picture；Recycling photographic head is clapped in real time Taking the photograph, the high clear colorful image of online acquisition product to be measured packaging, as testing image；
(2) based on SURF algorithm, testing image and standard picture are carried out images match；
(3) two images after coupling in step (2) are differed from shadow operation, obtain defect image；
(4) defect image is carried out feature extraction, obtain geometric properties and the color character of defect image；
(5) use RBF neural algorithm, using the characteristic vector of defect image as input value, make with the kind of defect For output valve, product surface of package defect is classified.
The present invention further preferably technical scheme is, in step (1), high clear colorful image is colored by 3CCD high definition face battle array Camera utilizes reflection source collection to obtain.
Preferably, the standard picture of step (1) is that the high clear colorful image that multiple no defective products are packed uses statistics flat All method synthesis.
Preferably, in step (2), the image matching method of testing image and standard picture comprises the steps:
A, in standard picture, a selected region, as template, uses SURF algorithm based on Color invariants, utilizes and schemes As the calculated Color invariants of colour information extracts the characteristic point of image；
After b, extraction characteristic point, the half-tone information in conjunction with image is characterized a generation Feature Descriptor；
C, employing Euclidean distance carry out similarity measurement, extract the characteristic point of coupling between two width images, after finding match point, On the basis of it, complete the registration of testing image and standard picture.
Preferably, the defect image in step (2) is corresponding poor pixel-by-pixel by the testing image after mating and standard picture Obtain.
Preferably, in step (4), the geometric properties of defect image include the length of defect area, width, circularity, area, Girth and direction；The color character of defect image includes color average and variance.
Preferably, in step (5) element number of the input layer of RBF neural by the defect map extracted in step (4) The geometric properties of picture and the total quantity of color character determine；The element number of output layer is true by the kind of product surface of package defect Fixed.
Beneficial effect: (1) present invention by Vision Builder for Automated Inspection carry out automatic defect detection with classification, can avoid artificial because of Element interference, is greatly lowered cost of labor, thus avoids the huge invisible costs such as training, the management that manual detection brings；This Invention uses SURF algorithm based on Color invariants to registrate testing image and standard picture, solves because of picture position The problem that the point that causes of deviation that exists, line defect pixel are blanked, has reached each picture of testing image and standard picture Element spatially one_to_one corresponding, and use SURF algorithm based on Color invariants to carry out registrating the face that can also utilize image Color information, using colouring information as feature, accelerates registration speed, improves detection efficiency；The present invention uses RBF neural, pin The learning capacity powerful to neutral net, identifies classification accurately to defect；
(2) Plays image of the present invention is the high clear colorful image employing statistical average method conjunction of multiple no defective products packaging Become, owing to each species diversity also can be there is between certified products, in order to make testing result effective, multiple certified products gathered are used system Meter averaging method synthesis is fabricated to standard picture, not only eliminates single image as the deviation caused by standard picture, the most also Represent the real information of image, improve defects detection and the degree of accuracy of classification；
(3) defect image in the present invention is obtained by difference shadow method, as long as guaranteeing that testing image and standard picture have reached essence Determining position, there is any difference, whether stain or cut between the two, can detect, Detection results carries at double Rise, and calculus of finite differences method is simple, easily realizes, meets the requirement that on-line detecting system code is succinct；
(4) when the present invention carrying out feature extraction to defect image, owing to defects detection problem is complex, each feature Between the degree of association relatively big, therefore the length of defect image, width, circularity, area, girth, direction, color average and variance are all made Being characterized extraction, obtain more feature as far as possible, the mathematical description of feature is more complete, and the information dropout of image is few, The classifying quality obtained under the conditions of same category device is more preferable.
Accompanying drawing explanation
Fig. 1 is the flow chart of the detection described in the embodiment of the present invention 1 and sorting technique；
Fig. 2 is RBF neural structure chart in the embodiment of the present invention 1；
Fig. 3 is the clustering algorithm figure of RBF neural in the embodiment of the present invention 1.
Detailed description of the invention
Below by accompanying drawing, technical solution of the present invention is described in detail, but protection scope of the present invention is not limited to Described embodiment.
Embodiment 1: a kind of product surface of package defects detection based on machine vision and sorting technique, specifically includes as follows Step:
(1) utilize high definition, high-speed camera head to gather the high clear colorful image of multiple no defective products packaging, use statistics flat All methods synthesize standard picture；Recycling photographic head captured in real-time, the high clear colorful image of online acquisition product to be measured packaging, make For testing image；
Statistical average method is the probability distribution according to each sampled pixel value, obtains assembly average as stencil value, if The sample image gathered is that N opens, and each sample image is expressed as: fi(x, y), i=0,1, L N-1, then each pixel value in template For:
(2) testing image and standard picture being registrated based on SURF algorithm, the concrete grammar of registration is: in standard picture A selected region, as template, uses SURF algorithm based on Color invariants, utilizes the calculated face of image color information Chromatic invariant extracts the characteristic point of image；Extract after characteristic point, be characterized in conjunction with the half-tone information of image and generate a feature and retouch State son；Use Euclidean distance to carry out similarity measurement, extract the characteristic point of coupling between two width images, after finding match point, with it On the basis of, complete the registration of testing image and standard picture,
For meeting human visual system and the approximate form of CIE-1964-XYZ benchmark, obtain (E, Eλ,Eλλ) and RGB color The relational expression of image:
E E λ E λ λ = 0.06 0.63 0.27 0.3 0.04 - 0.35 0.34 - 0.6 0.17 R G B
Color invariants is:
H i n v = E λ E λ λ ,
To RGB image, Color invariants H can be calculated by above-mentioned formulainv；
Standard picture and testing image are calculated Color invariants as input information, set up graphical rule space, then Use quick Hessian matrix to detect the extreme point on each tomographic image, in space any point (x, y), yardstick is σ, Hessian matrix is defined as:
H ( x , σ ) = L x x ( x , σ ) L x y ( x , σ ) L x y ( x , σ ) L y y ( x , σ ) ,
Here Lxx、Lxy、LyySecond dervative and input picture H for Gaussian functioninvThe result of convolution, obtains Hessian Determinant of a matrix is:
Det (H)=DxxDyy-(0.9Dxy)2,
Here Dxx、Dxy、DyyFor L during σ=1.2xx、Lxy、Lyy, extreme point that Hessian matrix is detected, set one Individual threshold value, when extreme point is more than threshold value, carries out non-maxima suppression to this extreme point in the three-dimensional neighborhood of 3 × 3 × 3, if Neighbour 26 near values are all big, then elect characteristic point as, carry out interpolation at metric space, obtain stable characteristic point position and place Scale-value.If extreme point is less than threshold value, then get rid of this point；
After obtaining characteristic point, it is first determined the principal direction of characteristic point, then coordinate axes is rotated to principal direction, by calculating Obtain description vectors, after normalization, form description of characteristic point, the similarity measurement employing Euclidean distance of two characteristic vectors:
D i j = ( Σ k = 0 k - n ( X i k - X j k ) 2 ) 1 / 2 ,
Wherein, XikRepresent the kth element that the ith feature point character pair in standard picture is vectorial, XjkRepresent template The kth element of the jth Feature point correspondence characteristic vector in image, n is characterized the dimension of vector, obtains one according to above formula Distance set, sets threshold value, when minimum range and time minimum range ratio are less than threshold value, then and the two Feature Points Matching；
(3) carry out after registration differing from shadow operation, obtain defect image, and defect image is processed, extract feature；
Defect image is carried out feature extraction, and owing to defects detection problem is complex, between each feature, the degree of association is relatively big, So obtaining more feature the most as far as possible, including geometric properties and color characteristic, geometric properties is defect area here Length, width, area, girth, circularity, direction, color characteristic is color average and variance, and the mathematical description of feature is the most complete, The information dropout of image is the fewest, and under the conditions of same category device, classifying quality is the best；
(4) use RBF neural algorithm that defect is classified, determine surface by geometric properties and color characteristic Defect is classified；
The characteristic vector that input is defect image of RBF neural, extracts the characteristic of defect image, including above-mentioned Geometric properties and color characteristic, using them as the input of network, so that it is determined that the unit number of input layer is 8, according to The identification requirement reached, effectively identifies the defect of product packaging, including aberration, stain, cutter silk, scumming, bites, flies Ink, the number therefore selecting output layer unit is 6, and the output of the most each unit represents a kind of defect type, the most just determines The input of network and output layer unit number, as in figure 2 it is shown, the output valve of network output layer is generally non-integer, in classification Time, in order to make output valve have meaning directly perceived, output is taken as integer, i.e. if yi≤ 0.1, then take yi=0；If yi>=0.9, then take yi=1, i=1, L, 6.If output vector is (0.998,0.02,0.004,0,0.015,0.1) ≈ (1,0,0,0,0,0), represent There is chromatic aberration defect；
Feature based on said extracted, packs defect image by RBF neural sorting algorithm to product and is analyzed, really Determining defect type, RBF neural training step is as follows: first determine hidden node in RBF net with clustering method (as shown in Figure 3) Data center, and the extension constant of hidden node is determined according to the distance between each data center, then use least square method meter Calculate the output weights of each hidden node,
The eigenvalue x that input is extractedi, i=1, L, 8, after determining data center, obtaining hidden layer output battle array is:
H ^ = [ h i j ] = Φ j ( | | x i - c j | | )
Then to all samples, network is output asOrderFor approximate error, by given output y= [y1,L,y6], minimizeOutput weightsAccording to the RBF nerve net inputted and train Network, output obtains defect classification.
Although as it has been described above, represented and described the present invention with reference to specific preferred embodiment, but it must not be explained For the restriction to the present invention self.Under the spirit and scope of the present invention premise defined without departing from claims, can be right Various changes can be made in the form and details for it.

Claims (7)

1. product surface of package defects detection based on machine vision and sorting technique, it is characterised in that comprise the steps:
(1) gather the high clear colorful image of no defective product packaging, make standard picture；Recycling photographic head captured in real-time, Line gathers the high clear colorful image of product to be measured packaging, as testing image；
(2) based on SURF algorithm, testing image and standard picture are carried out images match；
(3) two images after coupling in step (2) are differed from shadow operation, obtain defect image；
(4) defect image is carried out feature extraction, obtain geometric properties and the color character of defect image；
(5) RBF neural algorithm is used, using the characteristic vector of defect image as input value, using the kind of defect as defeated Go out value, product surface of package defect is classified.
Product surface of package defects detection based on machine vision the most according to claim 1 and sorting technique, its feature Being, in step (1), high clear colorful image is utilized reflection source collection to obtain by 3CCD high definition face battle array color camera.
Product surface of package defects detection based on machine vision the most according to claim 1 and sorting technique, its feature Being, the standard picture of step (1) is that the high clear colorful image of multiple no defective products packaging uses statistical average method synthesis.
Product surface of package defects detection based on machine vision the most according to claim 1 and sorting technique, its feature Being, in step (2), the image matching method of testing image and standard picture comprises the steps:
A, in standard picture, a selected region, as template, uses SURF algorithm based on Color invariants, utilizes image coloured silk The calculated Color invariants of color information extracts the characteristic point of image；
After b, extraction characteristic point, the half-tone information in conjunction with image is characterized a generation Feature Descriptor；
C, employing Euclidean distance carry out similarity measurement, extract the characteristic point of coupling between two width images, after finding match point, with it On the basis of, complete the registration of testing image and standard picture.
Product surface of package defects detection based on machine vision the most according to claim 1 and sorting technique, its feature Being, the defect image in step (2) is obtained by the testing image after mating and standard picture corresponding difference of making pixel-by-pixel.
Product surface of package defects detection based on machine vision the most according to claim 1 and sorting technique, its feature Being, in step (4), the geometric properties of defect image includes the length of defect area, width, circularity, area, girth and side To；The color character of defect image includes color average and variance.
Product surface of package defects detection based on machine vision the most according to claim 1 and sorting technique, its feature Being, in step (5), the element number of the input layer of RBF neural is special by the geometry of the defect image extracted in step (4) The total quantity of color character of seeking peace determines；The element number of output layer is determined by the kind of product surface of package defect.
CN201610579970.0A 2016-07-20 2016-07-20 Product surface of package defects detection based on machine vision and sorting technique Pending CN106204618A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610579970.0A CN106204618A (en) 2016-07-20 2016-07-20 Product surface of package defects detection based on machine vision and sorting technique

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610579970.0A CN106204618A (en) 2016-07-20 2016-07-20 Product surface of package defects detection based on machine vision and sorting technique

Publications (1)

Publication Number Publication Date
CN106204618A true CN106204618A (en) 2016-12-07

Family

ID=57491194

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610579970.0A Pending CN106204618A (en) 2016-07-20 2016-07-20 Product surface of package defects detection based on machine vision and sorting technique

Country Status (1)

Country Link
CN (1) CN106204618A (en)

Cited By (34)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106815830A (en) * 2016-12-13 2017-06-09 中国科学院自动化研究所 The defect inspection method of image
CN106855520A (en) * 2017-02-10 2017-06-16 南京航空航天大学 A kind of workpiece, defect detection method based on machine vision
CN107123108A (en) * 2017-03-30 2017-09-01 上海卫星装备研究所 Remote sensor light shield honeycomb defect automatic testing method based on machine vision
CN107948464A (en) * 2017-09-15 2018-04-20 兰州交通大学 A kind of geometric correction method and system of the laterally offset of printed matter detection image
CN108181316A (en) * 2018-03-13 2018-06-19 广西师范大学 A kind of bamboo strip defect detection method based on machine vision
CN108230327A (en) * 2016-12-14 2018-06-29 南京文采科技有限责任公司 A kind of packaging location based on MVP platforms and sort research universal method
CN108508844A (en) * 2018-04-17 2018-09-07 四川大学 Transferred product system and transfer method based on plane secondary light source and machine vision
CN109472783A (en) * 2018-10-31 2019-03-15 湘潭大学 A kind of nickel foam surface defect is extracted and classification method
CN109472769A (en) * 2018-09-26 2019-03-15 成都数之联科技有限公司 A kind of bad image defect detection method and system
CN109523505A (en) * 2018-09-18 2019-03-26 深圳市智信精密仪器有限公司 A kind of ceramic brick surface patterned defect detection method based on machine vision
CN109543565A (en) * 2018-11-05 2019-03-29 深圳和而泰数据资源与云技术有限公司 A kind of quantity determines method and device
CN109827971A (en) * 2019-03-19 2019-05-31 湖州灵粮生态农业有限公司 A kind of method of non-destructive testing fruit surface defect
CN109871895A (en) * 2019-02-22 2019-06-11 北京百度网讯科技有限公司 The defect inspection method and device of circuit board
CN109886325A (en) * 2019-02-01 2019-06-14 辽宁工程技术大学 A kind of stencil-chosen and acceleration matching process of non linear color space classification
CN109902633A (en) * 2019-03-04 2019-06-18 南京小网科技有限责任公司 Accident detection method and device based on the camera supervised video of fixed bit
CN109919907A (en) * 2019-01-23 2019-06-21 华灿光电（浙江）有限公司 The method and apparatus of LED epitaxial slice defects detection
CN110211102A (en) * 2019-05-23 2019-09-06 杭州良淋电子科技股份有限公司 A kind of racket wire break detection method of unmanned network renting and selling device
CN110252674A (en) * 2019-07-11 2019-09-20 重庆时代云英科技有限公司 A kind of method and system of the packing box quality testing based on machine vision
CN110378912A (en) * 2019-07-12 2019-10-25 深圳码隆科技有限公司 Wrap up detection method, device, computer readable storage medium and computer equipment
CN110570393A (en) * 2019-07-31 2019-12-13 华南理工大学 mobile phone glass cover plate window area defect detection method based on machine vision
CN110599484A (en) * 2019-09-19 2019-12-20 京东方科技集团股份有限公司 Defect detection method, device and storage medium
CN111189836A (en) * 2019-12-09 2020-05-22 杭州电子科技大学 Labview-based product defect detection method
CN111210412A (en) * 2019-12-31 2020-05-29 电子科技大学中山学院 Package detection method and device, electronic equipment and storage medium
CN111753877A (en) * 2020-05-19 2020-10-09 海克斯康制造智能技术(青岛)有限公司 Product quality detection method based on deep neural network transfer learning
CN111784672A (en) * 2020-06-30 2020-10-16 创新奇智(上海)科技有限公司 Defect detection method and device, electronic equipment and readable storage medium
CN112101836A (en) * 2020-03-06 2020-12-18 蒋梅 Big data storage node dynamic management system and corresponding terminal
CN112198161A (en) * 2020-10-10 2021-01-08 安徽和佳医疗用品科技有限公司 PVC gloves real-time detection system based on machine vision
CN112802022A (en) * 2021-04-14 2021-05-14 惠州高视科技有限公司 Method for intelligently detecting defective glass image, electronic device and storage medium
CN113109348A (en) * 2021-03-12 2021-07-13 华南理工大学 Paddle image transfer printing defect identification method based on machine vision
CN114418899A (en) * 2022-03-28 2022-04-29 深圳市嘉年印务有限公司 Self-adaptive repairing method and system for self-color printing and readable storage medium
CN114691722A (en) * 2022-03-25 2022-07-01 济宁市质量技术监督信息所 Method and system for quickly selecting package detection standard
CN117162665A (en) * 2023-11-03 2023-12-05 星璟材料科技(南通)有限公司 Printed matter printing production control system
CN117291918A (en) * 2023-11-24 2023-12-26 吉林大学 Automobile stamping part defect detection method based on three-dimensional point cloud
CN117764963A (en) * 2023-12-25 2024-03-26 深圳市恒顺亿包装材料有限公司 Packaging bag defect detection method, device, terminal and medium based on machine vision

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103593670A (en) * 2013-10-14 2014-02-19 浙江工业大学 Copper sheet and strip surface defect detection method based on-line sequential extreme learning machine
CN104568986A (en) * 2015-01-26 2015-04-29 中国科学院半导体研究所 Method for automatically detecting printing defects of remote controller panel based on SURF (Speed-Up Robust Feature) algorithm

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103593670A (en) * 2013-10-14 2014-02-19 浙江工业大学 Copper sheet and strip surface defect detection method based on-line sequential extreme learning machine
CN104568986A (en) * 2015-01-26 2015-04-29 中国科学院半导体研究所 Method for automatically detecting printing defects of remote controller panel based on SURF (Speed-Up Robust Feature) algorithm

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
余永维等: "基于深度学习网络的射线图像缺陷识别方法", 《仪器仪表学报》 *
王文举等: "彩色印刷品缺陷快速精确检测方法研究", 《包装工程》 *
韩英莉等: "一种带钢表面缺陷识别与分类的研究—基于混合加权特征和RBF网络的方法", 《计算机工程与应用》 *
高盛丰等: "基于颜色不变量和SURF的辞色图像凭借", 《计算机工程与应用》 *

Cited By (53)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106815830B (en) * 2016-12-13 2020-01-03 中国科学院自动化研究所 Image defect detection method
CN106815830A (en) * 2016-12-13 2017-06-09 中国科学院自动化研究所 The defect inspection method of image
CN108230327A (en) * 2016-12-14 2018-06-29 南京文采科技有限责任公司 A kind of packaging location based on MVP platforms and sort research universal method
CN106855520A (en) * 2017-02-10 2017-06-16 南京航空航天大学 A kind of workpiece, defect detection method based on machine vision
CN107123108A (en) * 2017-03-30 2017-09-01 上海卫星装备研究所 Remote sensor light shield honeycomb defect automatic testing method based on machine vision
CN107123108B (en) * 2017-03-30 2020-04-24 上海卫星装备研究所 Machine vision-based automatic detection method for honeycomb defects of remote sensor lens hood
CN107948464B (en) * 2017-09-15 2019-07-23 兰州交通大学 A kind of geometric correction method and system of the laterally offset of printed matter detection image
CN107948464A (en) * 2017-09-15 2018-04-20 兰州交通大学 A kind of geometric correction method and system of the laterally offset of printed matter detection image
CN108181316B (en) * 2018-03-13 2019-12-20 广西师范大学 Bamboo strip defect detection method based on machine vision
CN108181316A (en) * 2018-03-13 2018-06-19 广西师范大学 A kind of bamboo strip defect detection method based on machine vision
CN108508844A (en) * 2018-04-17 2018-09-07 四川大学 Transferred product system and transfer method based on plane secondary light source and machine vision
CN109523505B (en) * 2018-09-18 2021-06-11 深圳市智信精密仪器有限公司 Method for detecting pattern defects on surface of ceramic tile based on machine vision
CN109523505A (en) * 2018-09-18 2019-03-26 深圳市智信精密仪器有限公司 A kind of ceramic brick surface patterned defect detection method based on machine vision
CN109472769A (en) * 2018-09-26 2019-03-15 成都数之联科技有限公司 A kind of bad image defect detection method and system
CN109472783B (en) * 2018-10-31 2021-10-01 湘潭大学 Method for extracting and classifying surface defects of foamed nickel
CN109472783A (en) * 2018-10-31 2019-03-15 湘潭大学 A kind of nickel foam surface defect is extracted and classification method
CN109543565A (en) * 2018-11-05 2019-03-29 深圳和而泰数据资源与云技术有限公司 A kind of quantity determines method and device
CN109543565B (en) * 2018-11-05 2020-11-13 深圳和而泰数据资源与云技术有限公司 Quantity determination method and device
CN109919907A (en) * 2019-01-23 2019-06-21 华灿光电（浙江）有限公司 The method and apparatus of LED epitaxial slice defects detection
CN109886325A (en) * 2019-02-01 2019-06-14 辽宁工程技术大学 A kind of stencil-chosen and acceleration matching process of non linear color space classification
CN109886325B (en) * 2019-02-01 2022-11-29 辽宁工程技术大学 Template selection and accelerated matching method for nonlinear color space classification
CN109871895A (en) * 2019-02-22 2019-06-11 北京百度网讯科技有限公司 The defect inspection method and device of circuit board
CN109871895B (en) * 2019-02-22 2021-03-16 北京百度网讯科技有限公司 Method and device for detecting defects of circuit board
CN109902633A (en) * 2019-03-04 2019-06-18 南京小网科技有限责任公司 Accident detection method and device based on the camera supervised video of fixed bit
CN109827971A (en) * 2019-03-19 2019-05-31 湖州灵粮生态农业有限公司 A kind of method of non-destructive testing fruit surface defect
CN110211102A (en) * 2019-05-23 2019-09-06 杭州良淋电子科技股份有限公司 A kind of racket wire break detection method of unmanned network renting and selling device
CN110252674A (en) * 2019-07-11 2019-09-20 重庆时代云英科技有限公司 A kind of method and system of the packing box quality testing based on machine vision
CN110378912A (en) * 2019-07-12 2019-10-25 深圳码隆科技有限公司 Wrap up detection method, device, computer readable storage medium and computer equipment
CN110378912B (en) * 2019-07-12 2021-11-16 深圳码隆科技有限公司 Package detection method and device, computer readable storage medium and computer equipment
CN110570393A (en) * 2019-07-31 2019-12-13 华南理工大学 mobile phone glass cover plate window area defect detection method based on machine vision
CN110599484B (en) * 2019-09-19 2023-01-10 京东方科技集团股份有限公司 Defect detection method, device and storage medium
CN110599484A (en) * 2019-09-19 2019-12-20 京东方科技集团股份有限公司 Defect detection method, device and storage medium
CN111189836A (en) * 2019-12-09 2020-05-22 杭州电子科技大学 Labview-based product defect detection method
CN111210412A (en) * 2019-12-31 2020-05-29 电子科技大学中山学院 Package detection method and device, electronic equipment and storage medium
CN111210412B (en) * 2019-12-31 2024-03-15 电子科技大学中山学院 Packaging detection method and device, electronic equipment and storage medium
CN112101836B (en) * 2020-03-06 2021-07-09 江苏小梦科技有限公司 Big data storage node dynamic management system and corresponding terminal
CN112101836A (en) * 2020-03-06 2020-12-18 蒋梅 Big data storage node dynamic management system and corresponding terminal
CN111753877B (en) * 2020-05-19 2024-03-05 海克斯康制造智能技术(青岛)有限公司 Product quality detection method based on deep neural network migration learning
CN111753877A (en) * 2020-05-19 2020-10-09 海克斯康制造智能技术(青岛)有限公司 Product quality detection method based on deep neural network transfer learning
CN111784672A (en) * 2020-06-30 2020-10-16 创新奇智(上海)科技有限公司 Defect detection method and device, electronic equipment and readable storage medium
CN112198161A (en) * 2020-10-10 2021-01-08 安徽和佳医疗用品科技有限公司 PVC gloves real-time detection system based on machine vision
CN113109348A (en) * 2021-03-12 2021-07-13 华南理工大学 Paddle image transfer printing defect identification method based on machine vision
CN112802022B (en) * 2021-04-14 2021-10-29 高视科技(苏州)有限公司 Method for intelligently detecting defective glass image, electronic device and storage medium
CN112802022A (en) * 2021-04-14 2021-05-14 惠州高视科技有限公司 Method for intelligently detecting defective glass image, electronic device and storage medium
CN114691722B (en) * 2022-03-25 2024-05-03 济宁市质量技术监督信息所 Method and system for rapidly selecting package detection standard
CN114691722A (en) * 2022-03-25 2022-07-01 济宁市质量技术监督信息所 Method and system for quickly selecting package detection standard
CN114418899B (en) * 2022-03-28 2022-08-16 深圳市嘉年印务有限公司 Self-adaptive repairing method and system for self-color printing and readable storage medium
CN114418899A (en) * 2022-03-28 2022-04-29 深圳市嘉年印务有限公司 Self-adaptive repairing method and system for self-color printing and readable storage medium
CN117162665B (en) * 2023-11-03 2023-12-26 星璟材料科技(南通)有限公司 Printed matter printing production control system
CN117162665A (en) * 2023-11-03 2023-12-05 星璟材料科技(南通)有限公司 Printed matter printing production control system
CN117291918B (en) * 2023-11-24 2024-02-06 吉林大学 Automobile stamping part defect detection method based on three-dimensional point cloud
CN117291918A (en) * 2023-11-24 2023-12-26 吉林大学 Automobile stamping part defect detection method based on three-dimensional point cloud
CN117764963A (en) * 2023-12-25 2024-03-26 深圳市恒顺亿包装材料有限公司 Packaging bag defect detection method, device, terminal and medium based on machine vision

Similar Documents

Publication Publication Date Title
CN106204618A (en) Product surface of package defects detection based on machine vision and sorting technique
CN109977808B (en) Wafer surface defect mode detection and analysis method
CN108355987B (en) A kind of screen printing of battery quality determining method based on piecemeal template matching
CN108918536B (en) Tire mold surface character defect detection method, device, equipment and storage medium
CN108562589A (en) A method of magnetic circuit material surface defect is detected
CN109711288A (en) Remote sensing ship detecting method based on feature pyramid and distance restraint FCN
CN109087274A (en) Electronic device defect inspection method and device based on multidimensional fusion and semantic segmentation
CN106875373A (en) Mobile phone screen MURA defect inspection methods based on convolutional neural networks pruning algorithms
CN113920107A (en) Insulator damage detection method based on improved yolov5 algorithm
CN111985499B (en) High-precision bridge apparent disease identification method based on computer vision
CN109993710A (en) A kind of underwater picture denoising method based on generation confrontation network
CN109784204A (en) A kind of main carpopodium identification of stacking string class fruit for parallel robot and extracting method
CN112132196B (en) Cigarette case defect identification method combining deep learning and image processing
CN108428231A (en) A kind of multi-parameter Part Surface Roughness learning method based on random forest
CN111127417B (en) Printing defect detection method based on SIFT feature matching and SSD algorithm improvement
CN109919177A (en) Feature selection approach based on stratification depth network
CN109509171A (en) A kind of Fabric Defects Inspection detection method based on GMM and image pyramid
CN106355579A (en) Defect detecting method of cigarette carton surface wrinkles
CN109934221A (en) Power equipment based on attention mechanism automatically analyzes identification monitoring method and system
CN108932712A (en) A kind of rotor windings quality detecting system and method
CN103258218B (en) Masking-out detects skeleton generating method, device, defect inspection method and device
CN113205511A (en) Electronic component batch information detection method and system based on deep neural network
CN103034868B (en) Classification method aiming at small sample and high dimensional images
CN103544473B (en) A kind of electronic connector detection method based on machine vision
CN110956656A (en) Spindle positioning method based on depth target detection

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20161207