CN107886509A - A kind of image deflects recognition methods, electronic equipment, storage medium and system - Google Patents

A kind of image deflects recognition methods, electronic equipment, storage medium and system Download PDF

Info

Publication number
CN107886509A
CN107886509A CN201711191846.8A CN201711191846A CN107886509A CN 107886509 A CN107886509 A CN 107886509A CN 201711191846 A CN201711191846 A CN 201711191846A CN 107886509 A CN107886509 A CN 107886509A
Authority
CN
China
Prior art keywords
image
chaos
time sequence
pixel
histogram
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
CN201711191846.8A
Other languages
Chinese (zh)
Inventor
罗新斌
王勇
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.)
Sharp Ferroelectric Gas Science And Technology Ltd Of Suzhou Jade-Like Stone
Original Assignee
Sharp Ferroelectric Gas Science And Technology Ltd Of Suzhou Jade-Like Stone
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 Sharp Ferroelectric Gas Science And Technology Ltd Of Suzhou Jade-Like Stone filed Critical Sharp Ferroelectric Gas Science And Technology Ltd Of Suzhou Jade-Like Stone
Priority to CN201711191846.8A priority Critical patent/CN107886509A/en
Publication of CN107886509A publication Critical patent/CN107886509A/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
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/84—Systems specially adapted for particular applications
    • G01N21/88—Investigating the presence of flaws or contamination
    • G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/84—Systems specially adapted for particular applications
    • G01N21/88—Investigating the presence of flaws or contamination
    • G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
    • G01N2021/8887—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges based on image processing techniques
    • 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
    • G06T2207/10004—Still image; Photographic 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
    • G06T2207/20—Special algorithmic details
    • G06T2207/20081—Training; Learning

Landscapes

  • Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Artificial Intelligence (AREA)
  • Signal Processing (AREA)
  • Quality & Reliability (AREA)
  • Health & Medical Sciences (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Biochemistry (AREA)
  • General Health & Medical Sciences (AREA)
  • Immunology (AREA)
  • Pathology (AREA)
  • Image Analysis (AREA)

Abstract

The invention discloses a kind of image deflects recognition methods, electronic equipment, storage medium and system, this method includes：Often row pixel in image and each column pixel are obtained into the chaos characteristic of each chaos time sequence as chaos time sequence, computing；Chaos characteristic vector is established respectively to the chaos characteristic of each chaos time sequence；The chaos characteristic vector matrix of training sample is clustered using clustering algorithm, obtains code book；Each training image and the histogram of test image are calculated with bag of words according to code book；By multi-task learning method to training image and the histogram foundation group sparse model of test image；Utilize alternating direction Multiplier Algorithm calculating group sparse model；Defect image is classified using reconstructed error.Preferably surface defect can be identified by the present invention, and the degree of accuracy is high, can be applied in all kinds of civilian and military systems such as recognition of face, military target tracking and identifying system, has wide market prospects and application value.

Description

A kind of image deflects recognition methods, electronic equipment, storage medium and system
Technical field
The present invention relates to technical field of image processing, more particularly to a kind of image deflects recognition methods, electronic equipment, storage Medium and system.
Background technology
Defects detection important role in industrial automation, traditional detection method are mostly by physical measurement means To detect defect, image deflects detection is to reach testing goal by image, can be more directly perceived compared to traditional method See defect.Problems be present in existing image defect detection method：The defects of ignoring same kind is in surface image In similitude, cause detection efficiency low.
The content of the invention
In view of the shortcomings of the prior art, it is an object of the present invention to provide a kind of image deflects recognition methods, the side Method comprises the following steps：
Often row pixel in image and each column pixel are obtained into each chaotic time as chaos time sequence, computing The chaos characteristic of sequence；
Chaos characteristic vector is established respectively to the chaos characteristic of each chaos time sequence, obtains chaos characteristic vector Matrix；
The eigenvectors matrix of training sample is clustered using clustering algorithm, obtains code book；
The histogram of each training image is calculated with bag of words according to the code book, then calculates the straight of test image Fang Tu；
By multi-task learning method to the training image and the histogram foundation group sparse model of test image；
Described group of sparse model is calculated using alternating direction Multiplier Algorithm；
Defect image is classified using reconstructed error.
As a further improvement on the present invention, it is described that defect image is classified using reconstructed error, specifically include：Than Compared with the error between test sample and training sample, using nearest neighbouring rule, using error minimum as similar.
As a further improvement on the present invention, the chaos characteristic of the chaos time sequence includes：Embedded dimensions, when embedded Between postpone, box counting dimension, information dimension, pixel sequence average value and pixel sequence standard deviation.
As a further improvement on the present invention, the chaos characteristic vector is F=[τ, m, Di, Db, mean, std]；Wherein DiIt is information dimension, DbIt is box counting dimension, τ and m are that embedded delay and Embedded dimensions, mean represent the average value of pixel sequence respectively, Std represents the standard deviation of pixel sequence.
The second object of the present invention is to provide a kind of electronic equipment, including：Processor；Memory；And program, wherein Described program is stored in the memory, and is configured to by computing device, described program include being used for performing with Method described in upper any one.
The third object of the present invention is to provide a kind of computer-readable recording medium, is stored thereon with computer program, The computer program is executed by processor the method described in any of the above one.
The fourth object of the present invention is to provide a kind of image deflects identifying system, and the system is included with lower module：
Chaos time sequence computing module, using the often row pixel in image and each column pixel as chaotic time sequence Row, computing obtain the chaos characteristic of each chaos time sequence；
Chaos characteristic vector establishes module, and chaos characteristic is established respectively to the chaos characteristic of each chaos time sequence Vector, obtain chaos characteristic vector matrix；
Eigenvectors matrix cluster module, the eigenvectors matrix of training sample is clustered using clustering algorithm, obtained To code book；
Histogram calculation module, the histogram of each training image is calculated with bag of words according to the code book, then Calculate the histogram of test image；
Group sparse model establishes module, by multi-task learning method to the training image and the histogram of test image Foundation group sparse model；
Group sparse model computing module, described group of sparse model is calculated using alternating direction Multiplier Algorithm；
Defect image sort module, defect image is classified using reconstructed error.
As a further improvement on the present invention, it is described that defect image is classified using reconstructed error, specifically include：Than Compared with the error between test sample and training sample, using nearest neighbouring rule, using error minimum as similar.
As a further improvement on the present invention, the chaos characteristic of the chaos time sequence includes：Embedded dimensions, when embedded Between postpone, box counting dimension, information dimension, pixel sequence average value and pixel sequence standard deviation.
As a further improvement on the present invention, the chaos characteristic vector is F=[τ, m, Di, Db, mean, std]；Wherein DiIt is information dimension, DbIt is box counting dimension, τ and m are that embedded delay and Embedded dimensions, mean represent the average value of pixel sequence respectively, Std represents the standard deviation of pixel sequence.
Compared to prior art, the beneficial effects of the present invention are：The present invention proposes a kind of surface with robustness Defect identification method, the present invention simulate the pixel value changed over time with chaos time sequence, obtained by bag of words method straight Square figure carrys out the different image of comparison, and the relation between each training data is excavated with multi-task learning model, finally uses Optimization method calculates this model.Preferably surface defect can be identified by the present invention, compared to conventional method more It is accurate to add, and can be applied in all kinds of civilian and military systems such as recognition of face, military target tracking and identifying system, has wide Market prospects and application value.
Described above is only the general introduction of technical solution of the present invention, in order to better understand the technological means of the present invention, And can be practiced according to the content of specification, and in order to allow the above and other objects, features and advantages of the present invention can Become apparent, below especially exemplified by preferred embodiment, and coordinate accompanying drawing, describe in detail as follows.
Brief description of the drawings
Fig. 1 is the schematic diagram of image deflects recognition methods in the embodiment of the present invention；
Fig. 2 is the schematic diagram of image deflects identifying system in the embodiment of the present invention.
In figure：210th, chaos time sequence computing module；220th, chaos characteristic vector establishes module；230th, characteristic vector square Battle array cluster module；240th, histogram calculation module；250th, group sparse model establishes module；260th, sparse model computing module is organized； 270th, defect image sort module.
Embodiment
Below, with reference to accompanying drawing and embodiment, the present invention is described further：
As shown in figure 1, for the schematic diagram of image-recognizing method in the embodiment of the present invention, the described method comprises the following steps：
Step 110, the often row pixel in image and each column pixel are obtained each as chaos time sequence, computing The chaos characteristic of chaos time sequence；
Step 120, the chaos characteristic to each chaos time sequence establish chaos characteristic vector respectively, obtain chaos Eigenvectors matrix；
Step 130, using clustering algorithm the eigenvectors matrix of training sample is clustered, obtain code book；
Step 140, the histogram according to the code book with each training image of bag of words calculating, then calculate test The histogram of image；
Step 150, mould established to the histogram feature of the training image and test image by multi-task learning method Type；
Step 160, utilize the alternating direction Multiplier Algorithm calculating model.
As shown in Fig. 2 being the schematic diagram of image identification system in the embodiment of the present invention, the system is included with lower module：
Chaos time sequence computing module 210, using the often row pixel in image and each column pixel as chaotic time Sequence, computing obtain the chaos characteristic of each chaos time sequence；
Chaos characteristic vector establishes module 220, and chaos is established respectively to the chaos characteristic of each chaos time sequence Characteristic vector, obtain chaos characteristic vector matrix；
Eigenvectors matrix cluster module 230, the eigenvectors matrix of training sample is clustered using clustering algorithm, Obtain code book；
Histogram calculation module 240, the histogram of each training image is calculated with bag of words according to the code book, is connect The histogram for calculating test image；
Model building module 250, it is special to the training image and the histogram of test image by multi-task learning method Sign establishes model；
Model computation module 260, the model is calculated using alternating direction Multiplier Algorithm.
Compared to prior art, the beneficial effects of the present invention are：The present invention proposes a kind of surface with robustness Defect identification method, the present invention simulate the pixel value changed over time with chaos time sequence, obtained by bag of words method straight Square figure carrys out the different image of comparison, and the relation between each training data is excavated with multi-task learning model, finally uses Optimization method calculates this model.Preferably surface defect can be identified by the present invention, compared to conventional method more It is accurate to add, and can be applied in all kinds of civilian and military systems such as recognition of face, military target tracking and identifying system, has wide Market prospects and application value.
Above-mentioned embodiment is only the preferred embodiment of the present invention, it is impossible to the scope of protection of the invention is limited with this, The change and replacement for any unsubstantiality that those skilled in the art is done on the basis of the present invention belong to institute of the present invention Claimed scope.

Claims (10)

  1. A kind of 1. image deflects recognition methods, it is characterised in that including：
    Often row pixel in image and each column pixel are obtained into each chaos time sequence as chaos time sequence, computing Chaos characteristic；
    Chaos characteristic vector is established respectively to the chaos characteristic of each chaos time sequence, obtains chaos characteristic moment of a vector Battle array；
    The eigenvectors matrix of training sample is clustered using clustering algorithm, obtains code book；
    The histogram of each training image is calculated with bag of words according to the code book, then calculates the Nogata of test image Figure；
    By multi-task learning method to the training image and the histogram foundation group sparse model of test image；
    Described group of sparse model is calculated using alternating direction Multiplier Algorithm；
    Defect image is classified using reconstructed error.
  2. 2. image deflects recognition methods as claimed in claim 1, it is characterised in that described to be carried out using reconstructed error to defect image Classification, is specifically included：Compare the error between test sample and training sample, using nearest neighbouring rule, by the work that error is minimum To be similar.
  3. 3. image deflects recognition methods as claimed in claim 1, it is characterised in that the chaos characteristic bag of the chaos time sequence Include：Embedded dimensions, embedded time delay, box counting dimension, information dimension, pixel sequence average value and pixel sequence standard deviation.
  4. 4. image deflects recognition methods as claimed in claim 1, it is characterised in that the chaos characteristic vector is F=[τ, m, Di, Db, mean, std]；Wherein DiIt is information dimension, DbIt is box counting dimension, τ and m are embedded delay and Embedded dimensions respectively, and mean is represented The average value of pixel sequence, std represent the standard deviation of pixel sequence.
  5. 5. a kind of electronic equipment, it is characterised in that including：Processor；Memory；And program, wherein described program are stored In the memory, and it is configured to by computing device, described program includes requiring that 1-4 is any one for perform claim Method described in.
  6. 6. a kind of computer-readable recording medium, is stored thereon with computer program, it is characterised in that：The computer program quilt Method of the computing device as described in claim 1-4 any one.
  7. A kind of 7. image deflects identifying system, it is characterised in that including：
    Chaos time sequence computing module, using the often row pixel in image and each column pixel as chaos time sequence, transport Calculate the chaos characteristic for obtaining each chaos time sequence；
    Chaos characteristic vector establishes module, the chaos characteristic of each chaos time sequence is established respectively chaos characteristic to Amount, obtains chaos characteristic vector matrix；
    Eigenvectors matrix cluster module, the eigenvectors matrix of training sample is clustered using clustering algorithm, obtains generation Code book；
    Histogram calculation module, the histogram of each training image is calculated with bag of words according to the code book, is then calculated The histogram of test image；
    Group sparse model establishes module, and the training image and the histogram of test image are established by multi-task learning method Group sparse model；
    Group sparse model computing module, described group of sparse model is calculated using alternating direction Multiplier Algorithm；
    Defect image sort module, defect image is classified using reconstructed error.
  8. 8. image deflects recognition methods as claimed in claim 7, it is characterised in that described to be carried out using reconstructed error to defect image Classification, is specifically included：Compare the error between test sample and training sample, using nearest neighbouring rule, by the work that error is minimum To be similar.
  9. 9. image deflects recognition methods as claimed in claim 7, it is characterised in that the chaos characteristic bag of the chaos time sequence Include：Embedded dimensions, embedded time delay, box counting dimension, information dimension, pixel sequence average value and pixel sequence standard deviation.
  10. 10. image deflects recognition methods as claimed in claim 7, it is characterised in that the chaos characteristic vector is F=[τ, m, Di, Db, mean, std]；Wherein DiIt is information dimension, DbIt is box counting dimension, τ and m are embedded delay and Embedded dimensions respectively, and mean is represented The average value of pixel sequence, std represent the standard deviation of pixel sequence.
CN201711191846.8A 2017-11-24 2017-11-24 A kind of image deflects recognition methods, electronic equipment, storage medium and system Pending CN107886509A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201711191846.8A CN107886509A (en) 2017-11-24 2017-11-24 A kind of image deflects recognition methods, electronic equipment, storage medium and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201711191846.8A CN107886509A (en) 2017-11-24 2017-11-24 A kind of image deflects recognition methods, electronic equipment, storage medium and system

Publications (1)

Publication Number Publication Date
CN107886509A true CN107886509A (en) 2018-04-06

Family

ID=61774989

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201711191846.8A Pending CN107886509A (en) 2017-11-24 2017-11-24 A kind of image deflects recognition methods, electronic equipment, storage medium and system

Country Status (1)

Country Link
CN (1) CN107886509A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109522937A (en) * 2018-10-23 2019-03-26 北京市商汤科技开发有限公司 Image processing method and device, electronic equipment and storage medium
CN114140637A (en) * 2021-10-21 2022-03-04 阿里巴巴达摩院(杭州)科技有限公司 Image classification method, storage medium and electronic device
CN114943684A (en) * 2022-04-15 2022-08-26 上海波士内智能科技有限公司 Curved surface anomaly detection method by using confrontation to generate self-coding neural network

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102915451A (en) * 2012-10-18 2013-02-06 上海交通大学 Dynamic texture identification method based on chaos invariant
CN104361352A (en) * 2014-11-13 2015-02-18 东北林业大学 Solid wood panel defect separation method based on compressed sensing
US9165246B2 (en) * 2013-01-29 2015-10-20 Hewlett-Packard Development Company, L.P. Neuristor-based reservoir computing devices

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102915451A (en) * 2012-10-18 2013-02-06 上海交通大学 Dynamic texture identification method based on chaos invariant
US9165246B2 (en) * 2013-01-29 2015-10-20 Hewlett-Packard Development Company, L.P. Neuristor-based reservoir computing devices
CN104361352A (en) * 2014-11-13 2015-02-18 东北林业大学 Solid wood panel defect separation method based on compressed sensing

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
FABRICE AUZANNEAU等: "Chaos Time Domain Reflectometry for Online Defect Detection in Noisy Wired Networks", 《IEEE SENSORS JOURNAL》 *
佘骋南: "基于混沌特性的磁瓦表面缺陷视觉提取方法研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 *
王勇: "基于混沌特征向量的动态纹理识别", 《中国博士学位论文全文数据库 信息科技辑》 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109522937A (en) * 2018-10-23 2019-03-26 北京市商汤科技开发有限公司 Image processing method and device, electronic equipment and storage medium
CN114140637A (en) * 2021-10-21 2022-03-04 阿里巴巴达摩院(杭州)科技有限公司 Image classification method, storage medium and electronic device
CN114140637B (en) * 2021-10-21 2023-09-12 阿里巴巴达摩院(杭州)科技有限公司 Image classification method, storage medium and electronic device
CN114943684A (en) * 2022-04-15 2022-08-26 上海波士内智能科技有限公司 Curved surface anomaly detection method by using confrontation to generate self-coding neural network

Similar Documents

Publication Publication Date Title
Li et al. Automatic pavement crack detection by multi-scale image fusion
WO2020155518A1 (en) Object detection method and device, computer device and storage medium
CN103530590B (en) DPM Quick Response Code recognition system
CN108596010B (en) Implementation method of pedestrian re-identification system
CN107886509A (en) A kind of image deflects recognition methods, electronic equipment, storage medium and system
CN105005760A (en) Pedestrian re-identification method based on finite mixture model
CN114299150A (en) Depth 6D pose estimation network model and workpiece pose estimation method
CN107644203A (en) A kind of feature point detecting method of form adaptive classification
CN117671508B (en) SAR image-based high-steep side slope landslide detection method and system
CN105894014B (en) Abnormal behavior sequential detection method based on multi-factor inconsistency measurement
Li et al. Loop closure detection based on image semantic segmentation in indoor environment
CN116309882A (en) Tray detection and positioning method and system for unmanned forklift application
Li et al. Robot vision model based on multi-neural network fusion
CN113743382A (en) Shelf display detection method, device and system
WO2020233414A1 (en) Object recognition method and apparatus, and vehicle
CN108830281B (en) Repeated image matching method based on local change detection and spatial weighting
CN113989632A (en) Bridge detection method and device for remote sensing image, electronic equipment and storage medium
CN114758259B (en) Package detection method and system based on X-ray object image recognition
Lu et al. Design of Mixed Reality Assembly System Based on Improved ORB Algorithm
Rao et al. Research on Feature Extraction and Matching Algorithm in Visual SLAM
Wang et al. A-BEBLID: A Hybrid Image Registration Method for Lithium-Ion Battery Cover Screen Printing
Zhao et al. Modern Craft Product Design Using Digital Technology Combined with Visual Sensing System in Complex Environment
CN205068476U (en) Automatic article management system based on image matching
CN111079775B (en) Real-time tracking method for combined regional constraint learning
Zeno et al. Event recognition on images using support vector machine and multi-level histograms of local patterns

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
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20180406