CN105869170A - Identification and classification method for workpiece surface texture image - Google Patents
Identification and classification method for workpiece surface texture image Download PDFInfo
- Publication number
- CN105869170A CN105869170A CN201610229339.8A CN201610229339A CN105869170A CN 105869170 A CN105869170 A CN 105869170A CN 201610229339 A CN201610229339 A CN 201610229339A CN 105869170 A CN105869170 A CN 105869170A
- Authority
- CN
- China
- Prior art keywords
- workpiece surface
- surface texture
- image
- texture image
- identification
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
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
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
-
- 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
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/40—Analysis of texture
-
- 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
- G06T2207/30164—Workpiece; Machine component
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an identification and classification method for a workpiece surface texture image. Particularly, the identification and classification method comprises the steps of a first step, extracting texture characteristic of the workpiece surface texture image; a second step, extracting shape characteristic of the workpiece surface texture image; a third step, integrating a texture characteristic vector and a shape characteristic vector of the workpiece surface texture image; and a fourth step, performing identification and classification on the workpiece surface texture image by means of a support vector machine method. According to the identification and classification method, through extracting two characteristics of the workpiece surface texture image, and performing training and identification, characteristic extracting performance for invariants in expansion Shearlet transform and Krawchouk moment and identification performance of the support vector machine are sufficiently utilized. The identification and classification method has advantages of wholly describing the texture characteristic of the workpiece surface image, improving accuracy and robustness of the identification and classification method for the workpiece surface texture image, finishing identification and classification on the workpiece surface image by means of the support vector machine which is optimized by a chaotic glowworm swarm algorithm.
Description
Technical field
The present invention relates to technical field of image processing, be specifically related to the identification classification side of a kind of workpiece surface texture image
Method.
Background technology
Along with computer technology and the fast development of image processing techniques, production automation degree, machining accuracy are increasingly
Height, cutting-tool wear state based on workpiece surface texture image detection is widely used in automated production processing, at present to work
The identification of part skin texture images is the important content of tool wear monitoring based on surface of the work image.
The recognition methods of existing workpiece texture image have method based on neutral net, method based on fuzzy clustering,
Method based on genetic algorithm etc., but due in actual workpiece surface texture image acquisition procedures, by noise, illumination etc. outside
The interference of boundary's enchancement factor makes the original texture image obtained of low quality so that existing recognition methods often identifies classification
Accuracy rate is low.
Summary of the invention
For the problems referred to above, the invention provides the method for identifying and classifying of a kind of workpiece surface texture image, its purpose exists
In: improve the accuracy of method for identifying and classifying of workpiece surface texture image, robustness.
The technical solution of the present invention:
A kind of method for identifying and classifying of workpiece surface texture image, concrete identification classifying step is:
Step 1: the texture feature extraction of workpiece surface texture image；
Step 2: the Shape Feature Extraction of workpiece surface texture image；
Step 3: texture feature vector and the shape eigenvectors of workpiece surface texture image are comprehensive;
Step 4: utilize the method for SVMs to carry out the identification classification of workpiece surface texture image.
Inventive feature also resides in,
In step 1, the process of the texture feature extraction of image is:
Step 1.1: first willThe workpiece surface texture image of size is extended Shearlet and decomposes, and obtains m son
Band image (m level cone sub-band images and m vertically bore sub-band images), can try to achieve different directions i and the level of different scale j
The Shearlet conversion coefficient of cone sub-band imagesWith the vertical Shearlet conversion coefficient boring sub-band images。
Step 1.2: first calculate the mean value of square of sub-band images, i.e.
(1)
In formulaRepresent the Energy distribution average of sub-band images；WithRepresent line number and the columns of sub-band images respectively；Represent horizontal awl band image coefficientOr vertical awl band image coefficient。
Then yardstick is calculatedUnder average energy value sum, the Energy distribution average after being weighted, then surface of the work
The level cone texture feature vector of texture image is 。
Step 1.3: according to step 1.2, the vertical cone texture feature vector that can try to achieve workpiece surface texture image equally is, then the texture feature vector of view picture workpiece surface texture image is。
In step 2, the process of the Shape Feature Extraction of image:
If image size is, as the shape facility of workpiece surface texture imageRank Krawtchouk square
Invariant is defined as:
(2)
In formula,,,。
Then the shape eigenvectors of view picture workpiece surface texture image is。
In step 3, texture feature vector and the comprehensive process of shape eigenvectors:
Step 3.1: by formula (3) by texture feature vectorElement normalization:
(3)
Obtain the texture feature vector of normalized workpiece surface texture image。
Step 3.2: according to the normalized process of step 3.1, can obtain normalized workpiece surface texture figure
The shape eigenvectors of picture。
Step 3.3: the texture feature vector of comprehensive workpiece surface texture imageAnd shape eigenvectors, constitute view picture
The multi-feature vector of imageFor。
In step 4, the identification assorting process of workpiece surface texture image:
First the training sample of workpiece surface texture image and test sample are performed step 1 to 3, respectively obtain training sample and
Test sample characteristic vector, is then input in the SVMs that chaos firefly group optimizes, it is thus achieved that optimum parameter is arranged.
Finally to workpiece surface texture image execution step 1 to be identified to 3, by defeated for the workpiece surface texture image feature vector obtained
Enter SVMs, finally give identification and the classification results of workpiece surface texture image.
Beneficial effects of the present invention:
The present invention, by surface of the work literary composition texture image is carried out two kinds of feature extractions, is then trained and identifies, fully profit
With extension Shearlet conversion and the feature extraction performance of Krawchouk moment invariants and the recognition performance of SVMs, energy
It is more fully described by the textural characteristics of surface of the work image, improves the standard of the method for identifying and classifying of workpiece surface texture image
Really property, robustness, uses the SVMs of chaos firefly group's algorithm optimization to complete surface of the work image recognition classification, improves
Discrimination.
Accompanying drawing explanation
Fig. 1: the flow chart of the method for identifying and classifying of workpiece surface texture image of the present invention.
Detailed description of the invention
With embodiment, the present invention is described further below in conjunction with the accompanying drawings:
The present invention is that the flow process of the method for identifying and classifying of a kind of workpiece surface texture image is as it is shown in figure 1, specifically comprise the following steps that
Step 1: the texture feature extraction of image:
The present invention uses extension Shearlet conversion to obtain the texture feature vector of workpiece surface texture image.First will
The workpiece surface texture image of size is extended Shearlet and decomposes, and obtains m sub-band images (m level cone sub-band images
Sub-band images is vertically bored with m), the Shearlet conversion of the level cone sub-band images of different directions i and different scale j can be tried to achieve
CoefficientWith the vertical Shearlet conversion coefficient boring sub-band images.Then the mean value of square of sub-band images is calculated,
I.e.
(1)
In formulaRepresent the Energy distribution average of sub-band images；WithRepresent line number and the columns of sub-band images respectively；Represent horizontal awl band image coefficientOr vertical awl band image coefficient。
Then yardstick is calculatedUnder average energy value sum, the Energy distribution average after being weighted, then surface of the work
The level cone texture feature vector of texture image is.That can try to achieve workpiece surface texture image equally vertically bores line
Reason characteristic vector is, then the texture feature vector of view picture workpiece surface texture image is。
Step 2: the Shape Feature Extraction of image:
The present invention utilizes Krawtchouk moment invariants to be calculated the shape eigenvectors of workpiece surface texture image.If image
Size is, as the shape facility of workpiece surface texture imageRank Krawtchouk moment invariants definition
For:
(2)
In formula,,,
。
Then the shape eigenvectors of view picture workpiece surface texture image is。
Step 3: texture feature vector and the comprehensive process of shape eigenvectors:
By formula (3) by texture feature vectorElement normalization,
(3)
Obtain the texture feature vector of normalized workpiece surface texture image。
Equally obtain the shape eigenvectors of normalized workpiece surface texture image
.The texture feature vector of comprehensive workpiece surface texture imageAnd shape eigenvectors, constitute the comprehensive characteristics of entire image
VectorFor。
Step 4: the identification assorting process of workpiece surface texture image:
The present invention uses the SVMs of chaos firefly group's algorithm optimization to complete the identification classification of workpiece texture image.First
Training sample and test sample execution step 1 to workpiece surface texture image, to 3, respectively obtain training sample and test sample
Characteristic vector, is then input in the SVMs that chaos firefly group optimizes, it is thus achieved that optimum parameter is arranged.Finally treat
Identify workpiece surface texture image perform step 1 to 3, by obtain workpiece surface texture image feature vector input support to
Amount machine, finally gives identification and the classification results of workpiece surface texture image.
The test result test experiments result of the present invention is as shown in table 1 below, based on Gabor wavelet and total knowledge of SVM method
Not rate is 87.8%, and total discrimination based on gray level co-occurrence matrixes and BP neural net method is 84.4%, total identification of the present invention
Rate is 96.7%.
The comparison of 13 kinds of method recognition results of table
The present invention uses extension Shearlet conversion to extract the textural characteristics of surface of the work image, and extension Shearlet converts not only
There is directional sensitivity, the feature that spatially localized, parabola is sized and optimum is sparse, and its shearing wave is to each chi
Degree, direction and position all can realize preferably location, therefore effectively extract the texture information of workpiece surface texture image.
Krawtchouk square is discrete orthogonal moments, and the error overcoming the continuity moments such as Zernike square, Zernike pseudo-matrix is big, computationally intensive
Shortcoming, and Krawtchouk moment invariants has translation, rotates and scale invariability, can Efficient Characterization workpiece surface texture figure
The shape facility of picture.Extension Shearlet conversion and Krawtchouk moment invariants are used in combination, can be more fully described by
The textural characteristics of surface of the work image.The SVMs using chaos firefly group's algorithm optimization completes surface of the work image to be known
Do not classify, improve discrimination.The present invention uses extension Shearlet conversion, Krawchouk moment invariants and chaos firefly
The workpiece surface texture image recognition sorting technique of Support Vector Machines Optimized with based on Gabor wavelet and SVM method, based on gray scale
Co-occurrence matrix and the contrast identifying classification of BP neural net method, as shown in table 1.
The above, the only detailed description of the invention of the present invention, but protection scope of the present invention is not limited thereto, and any
Those familiar with the art in the technical scope that the invention discloses, the change that can readily occur in or replacement, all answer
Contain within protection scope of the present invention.
Claims (5)
1. the method for identifying and classifying of a workpiece surface texture image, it is characterised in that: comprise the steps:
Step 1: workpiece surface texture image is carried out texture feature extraction；
Step 2: workpiece surface texture image is carried out Shape Feature Extraction；
Step 3: by comprehensive to texture feature vector and the shape eigenvectors of workpiece surface texture image；
Step 4: the identification utilizing the method for SVMs to carry out workpiece surface texture image is classified.
The method for identifying and classifying of a kind of workpiece surface texture image the most according to claim 1, it is characterised in that: described step
The process of the texture feature extraction of the image in rapid 1 is:
The first step, willThe workpiece surface texture image of size is extended Shearlet and decomposes, and obtains m sub-band images
(m level cone sub-band images and m vertically bore sub-band images), can try to achieve the horizontal awl band of different directions i and different scale j
The Shearlet conversion coefficient of imageWith the vertical Shearlet conversion coefficient boring sub-band images；
Second step, calculates the mean value of square of sub-band images, i.e.
Wherein,Represent the Energy distribution average of sub-band images,WithRepresent line number and the columns of sub-band images respectively,Represent horizontal awl band image coefficientOr vertical awl band image coefficient；
Then yardstick is calculatedUnder average energy value sum, the Energy distribution average after being weighted, then workpiece surface texture
The level cone texture feature vector of image is；
3rd step, the vertical cone texture feature vector that can try to achieve workpiece surface texture image according to second step equally is, the most whole
The texture feature vector of width workpiece surface texture image is。
The method for identifying and classifying of a kind of workpiece surface texture image the most according to claim 1, it is characterised in that: described step
The process of the Shape Feature Extraction of the image in rapid 2 is:
If image size is, as the shape facility of workpiece surface texture imageRank Krawtchouk square is not
Variable-definition is:
Wherein,,,；
Then the shape eigenvectors of view picture workpiece surface texture image is。
The method for identifying and classifying of a kind of workpiece surface texture image the most according to claim 1, it is characterised in that described step
Texture feature vector in rapid 3 and the comprehensive process of shape eigenvectors:
The first step: by texture feature vectorElement normalization:
Obtain the texture feature vector of normalized workpiece surface texture image；
Second step: according to the normalized process of the first step, the shape of normalized workpiece surface texture image can be obtained
Shape characteristic vector；
3rd step: the texture feature vector of comprehensive workpiece surface texture imageAnd shape eigenvectors, constitute entire image
Multi-feature vectorFor。
The method for identifying and classifying of a kind of workpiece surface texture image the most according to claim 1, it is characterised in that: described step
The identification assorting process of the workpiece surface texture image in rapid 4 is:
First the training sample of workpiece surface texture image and test sample are performed step 1 to 3, respectively obtain training sample and
Test sample characteristic vector；
It is then input in the SVMs that chaos firefly group optimizes, it is thus achieved that optimum parameter is arranged；
Finally workpiece surface texture image to be identified is performed step 1 to 3, by the workpiece surface texture characteristics of image that obtains to
Amount input SVMs, finally gives identification and the classification results of workpiece surface texture image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610229339.8A CN105869170A (en) | 2016-04-13 | 2016-04-13 | Identification and classification method for workpiece surface texture image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610229339.8A CN105869170A (en) | 2016-04-13 | 2016-04-13 | Identification and classification method for workpiece surface texture image |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105869170A true CN105869170A (en) | 2016-08-17 |
Family
ID=56636919
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610229339.8A Pending CN105869170A (en) | 2016-04-13 | 2016-04-13 | Identification and classification method for workpiece surface texture image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105869170A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106485278A (en) * | 2016-10-13 | 2017-03-08 | 河南科技大学 | A kind of image texture sorting technique based on shearing wave and gauss hybrid models |
| CN108363942A (en) * | 2017-12-26 | 2018-08-03 | 新智数字科技有限公司 | A kind of tool recognizing method, apparatus based on multi-feature fusion and equipment |
| CN112052912A (en) * | 2020-09-23 | 2020-12-08 | 同济大学 | Intelligent flame combustion state identification method for fire-fighting robot |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103544499A (en) * | 2013-10-12 | 2014-01-29 | 江南大学 | Method for reducing dimensions of texture features for surface defect detection on basis of machine vision |
| CN104156726A (en) * | 2014-08-19 | 2014-11-19 | 大连理工大学 | Workpiece recognition method based on geometric shape feature and device thereof |
-
2016
- 2016-04-13 CN CN201610229339.8A patent/CN105869170A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103544499A (en) * | 2013-10-12 | 2014-01-29 | 江南大学 | Method for reducing dimensions of texture features for surface defect detection on basis of machine vision |
| CN104156726A (en) * | 2014-08-19 | 2014-11-19 | 大连理工大学 | Workpiece recognition method based on geometric shape feature and device thereof |
Non-Patent Citations (2)
| Title |
|---|
| 吴一全 等: "基于扩展Shearlet变换、Krawtchouk矩和SVM的储粮害虫分类", 《中国粮油学报》 * |
| 吴诗婳 等: "基于Shearlet变换和Krawtchouk矩不变量的河流SAR图像分割", 《应用科学学报》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106485278A (en) * | 2016-10-13 | 2017-03-08 | 河南科技大学 | A kind of image texture sorting technique based on shearing wave and gauss hybrid models |
| CN108363942A (en) * | 2017-12-26 | 2018-08-03 | 新智数字科技有限公司 | A kind of tool recognizing method, apparatus based on multi-feature fusion and equipment |
| CN112052912A (en) * | 2020-09-23 | 2020-12-08 | 同济大学 | Intelligent flame combustion state identification method for fire-fighting robot |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111292305B (en) | Improved YOLO-V3 metal processing surface defect detection method | |
| CN108765412B (en) | Strip steel surface defect classification method | |
| CN107230203B (en) | Casting defect identification method based on human eye visual attention mechanism | |
| Xie | A review of recent advances in surface defect detection using texture analysis techniques | |
| Hassanin et al. | A real-time approach for automatic defect detection from PCBs based on SURF features and morphological operations | |
| CN107767387B (en) | Contour detection method based on variable receptive field scale global modulation | |
| CN107085846B (en) | Workpiece surface defect image identification method | |
| US20160196467A1 (en) | Three-Dimensional Face Recognition Device Based on Three Dimensional Point Cloud and Three-Dimensional Face Recognition Method Based on Three-Dimensional Point Cloud | |
| Pramunendar et al. | A classification method of coconut wood quality based on Gray Level Co-occurrence matrices | |
| CN103996052B (en) | Three-dimensional face gender classification method based on three-dimensional point cloud | |
| CN105069447B (en) | A kind of recognition methods of human face expression | |
| CN102722708B (en) | Method and device for classifying sheet media | |
| Aptoula et al. | Morphological features for leaf based plant recognition | |
| CN102508917B (en) | Multi-dimensional object robust high-speed retrieval and positioning method for some feature images | |
| CN107886539B (en) | High-precision gear visual detection method in industrial scene | |
| CN108171102A (en) | A kind of part method for quickly identifying of view-based access control model | |
| CN104680550A (en) | Method for detecting defect on surface of bearing by image feature points | |
| CN101996328A (en) | Wood identification method | |
| CN108090492B (en) | Contour detection method based on scale clue suppression | |
| CN102136074B (en) | Man-machine interface (MMI) based wood image texture analyzing and identifying method | |
| CN105869170A (en) | Identification and classification method for workpiece surface texture image | |
| CN104537363A (en) | Full-automatic adjustable cupboard leg assembly control method based on visual inspection system | |
| Meruliya et al. | Image processing for fruit shape and texture feature extraction-review | |
| CN112132153A (en) | Tomato fruit identification method and system based on clustering and morphological processing | |
| Narayan et al. | An optimal feature subset selection using GA for leaf classification |
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: 20160817 |