CN105678341B - A kind of woollen and cashmere recognizer based on Gabor wavelet analysis - Google Patents
A kind of woollen and cashmere recognizer based on Gabor wavelet analysis Download PDFInfo
- Publication number
- CN105678341B CN105678341B CN201610091974.4A CN201610091974A CN105678341B CN 105678341 B CN105678341 B CN 105678341B CN 201610091974 A CN201610091974 A CN 201610091974A CN 105678341 B CN105678341 B CN 105678341B
- Authority
- CN
- China
- Prior art keywords
- image
- cashmere
- gabor
- woollen
- feature
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
-
- 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
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
Abstract
A kind of woollen and cashmere recognizer based on Gabor wavelet analysis, including online identification process and model learning stream, the online recognition flow include the following steps：(1) capture is carried out to Cashmere and Woolens fiber；(2) it uses Gaussian filter to carry out smothing filtering to image, image gray levels is adjusted and realize image enhancement；(3) using the extraction image object of edge detection and profile based on canny；(4) Gabor characteristic is extracted；(5) result of calculation；The model learning flow, includes the following steps：(1) database of a large amount of woollen and cashmere accumulated；(2) type and the location of hard objectives fiber are manually marked；(3) fibre image in database is pre-processed, feature extraction, the step is consistent with (2) (4) step of online recognition flow；(4) learning process takes the SVM classifier of two classification.The present invention can reach discrimination height, and recognition speed is fast, and recognition result accuracy rate is high, the advantageous effect of strong applicability.
Description
Technical field
The invention belongs to woollen and cashmere identification technology field more particularly to a kind of wool sheep based on Gabor wavelet analysis
Suede recognizer.
Background technology
Cashmere fiber is elongated, uniform, soft, has the characteristics that softly to slide with its manufactured textile and warm up, is high-grade dress ornament
The first choice faced.Due to its yield rareness, on the high side, manufacturing enterprise is often carried out with the Cashmere and Woolens of different proportion blended.Sheep
Hair belongs to natural protein fibre with cashmere, and structure and form are all very close, accurately carries out sentencing for kinds of fibers
Disconnected is very difficult.
Currently used fibre identification method is microscopic method.Testing staff is by observing wool sheep under the microscope
The features such as the scale shape and grain details of suede carry out qualitative classification according to its personal experience to the ingredient of Cashmere and Woolens, this
Kind mode not only takes time and effort, and subjectivity is big, and the consistency of measurement is also poor.
Reference standard of the present invention：1,2,3, propose a kind of intelligent full automatic Cashmere and Woolens recognition methods, first with
The Image Acquisition that microscope is digitized Cashmere and Woolens with CCD recycles and carries out different scale to Cashmere and Woolens image
Under wavelet convolution extract feature, and build sorter model using SVM, realize and the intelligent classification of Cashmere and Woolens is identified.
Gabor wavelet and the visual stimulus response of simple cell in human visual system are closely similar.It is in extraction target
Local space and frequency-domain information in terms of have good characteristic.Gabor wavelet is capable of providing the edge sensitive of image
Good set direction and scale selection characteristic, and it is insensitive for illumination variation, it is capable of providing good to illumination variation
Adaptability.Two-Dimensional Gabor Wavelets transformation is that the important tool of signal analysis and processing is carried out in time-frequency domain, and transformation coefficient has good
Good visual characteristic and Biological background, is therefore widely used in the fields such as image procossing, pattern-recognition.With traditional Fourier
Leaf transformation is compared, and Gabor wavelet transformation has good Time-Frequency Localization characteristic.
Invention content
The present invention provides a kind of woollen and cashmere recognizer analyzed based on Gabor wavelet, can reach discrimination height, identification
Speed is fast, and recognition result accuracy rate is high, the advantageous effect of strong applicability.
Technical problem solved by the invention is realized using following technical scheme：The present invention provides a kind of based on Gabor
The woollen and cashmere recognizer of wavelet analysis, including online identification process and model learning flow：
The online recognition flow carries out qualitative analysis to the fibre image acquired in real time, includes the following steps：
(1) acquisition of image coordinates Olympus CX41 biomicroscopes, to cashmere using 3,000,000 pixel technical grade ccd
Wool fiber carries out capture；
(2) it pre-processes：A carries out smothing filtering using Gaussian filter to image, to remove the noise in image, Gauss filter
Wave device is a kind of low-pass filter, and process can be using Formal Representation as input picture I's (x, y) and gaussian kernel function G (x, y)
Convolution：
S (x, y)=I (x, y) × G (x, y；σ) wherein
B adjusts image gray levels and realizes image enhancement, if data xijIt is the i row j column elements in image X, maxx, minx
It is maximum, the minimum value in X respectively；
(3) image object extracts：Using the extraction of edge detection and profile based on canny, canny edge detections are one
A multistage edge detection algorithm；Its basic step mainly has：A obtains x, the gradient of y, and the non-maximum values of b inhibit, c Edge tracks, this
In directly use opencv inside canny operators；
(4) Gabor wavelet feature is extracted：Divide 5 40, the direction of scale 8 Feature Descriptors；Using PCA dimensionality reductions to 100
Dimension；
Gabor functions can extract relevant feature on frequency domain different scale, different directions；Two-dimensional Gabor function can
To be expressed as：
Wherein：
The Gabor bases for choosing one group of different scale and direction carry out convolution to input fibre image, specifically choose 4 scales
(v=0,1 ..., 3), 8 directions (i.e. K=8, u=0,1 ..., 7), totally 32 Gabor kernel functions；
(5) feature vector is inputted into classifier calculated result；Each fibre image will be inputted, and to be uniformly scaled 48*48 big
It is small, the Gabor characteristic collection of illustrative plates under 32 scales is then extracted, extracts 48*48*32=73728 dimensional vectors altogether；Further we adopt
Take PCA methods that original high dimensional data is down to 100 dimensions, the input feature value as SVM；
The model learning flow is to obtain a grader, takes a kind of sorter model based on SVM, including
Following steps：
(1) premise of model learning is the database of a large amount of woollen and cashmere of accumulation；
(2) herein on basis, by the way of manually marking, make the type of machine hard objectives fiber and residing position
It sets, is a kind of supervised learning mode；
(3) fibre image in database is pre-processed, feature extraction, (2) (4) of the step and online recognition flow
Step is consistent；
(4) learning process takes the SVM classifier of two classification, and is trained using the libSVM to increase income, chooses RBF diameters
To base kernel function, iterations are set as 100000 times, iteration ends deviation 0.001.
Beneficial effects of the present invention are：
The present invention is high to the discrimination of Cashmere and Woolens, and recognition speed is fast, recognition result accuracy rate height and strong applicability.
Description of the drawings
Fig. 1 is the flow diagram of the present invention.
Specific implementation mode
Embodiment：
The present invention includes online identification process and model learning flow：
The online recognition flow carries out qualitative analysis to the fibre image acquired in real time, includes the following steps：
(1) acquisition of image coordinates Olympus CX41 biomicroscopes, to cashmere using 3,000,000 pixel technical grade ccd
Wool fiber carries out capture；
(2) it pre-processes：A carries out smothing filtering using Gaussian filter to image, to remove the noise in image, Gauss filter
Wave device is a kind of low-pass filter, and process can be using Formal Representation as input picture I's (x, y) and gaussian kernel function G (x, y)
Convolution：
S (x, y)=I (x, y) × G (x, y；σ) wherein
B adjusts image gray levels and realizes image enhancement, if data xijIt is the i row j column elements in image X, maxx, minx
It is maximum, the minimum value in X respectively；
(3) image object extracts：Using the extraction of edge detection and profile based on canny, canny edge detections are one
A multistage edge detection algorithm；Its basic step mainly has：A obtains x, the gradient of y, and the non-maximum values of b inhibit, c Edge tracks, this
In directly use opencv inside canny operators；
(4) Gabor wavelet feature is extracted：Divide 5 40, the direction of scale 8 Feature Descriptors；Using PCA dimensionality reductions to 100
Dimension；
Gabor functions can extract relevant feature on frequency domain different scale, different directions；Two-dimensional Gabor function can
To be expressed as：
Wherein：
The Gabor bases for choosing one group of different scale and direction carry out convolution to input fibre image, specifically choose 4 scales
(v=0,1 ..., 3), 8 directions (i.e. K=8, u=0,1 ..., 7), totally 32 Gabor kernel functions；
(5) feature vector is inputted into classifier calculated result；Each fibre image will be inputted, and to be uniformly scaled 48*48 big
It is small, the Gabor characteristic collection of illustrative plates under 32 scales is then extracted, extracts 48*48*32=73728 dimensional vectors altogether；Further we adopt
Take PCA methods that original high dimensional data is down to 100 dimensions, the input feature value as SVM；
The model learning flow is to obtain a grader, takes a kind of sorter model based on SVM, including
Following steps：
(1) premise of model learning is the database of a large amount of woollen and cashmere of accumulation；
(2) herein on basis, by the way of manually marking, make the type of machine hard objectives fiber and residing position
It sets, is a kind of supervised learning mode；
(3) fibre image in database is pre-processed, feature extraction, (2) (4) of the step and online recognition flow
Step is consistent；
(4) learning process takes the SVM classifier of two classification, and is trained using the libSVM to increase income, chooses RBF diameters
To base kernel function, iterations are set as 100000 times, iteration ends deviation 0.001.
The present invention is described in detail above by embodiment, but the content is only the preferable implementation of the present invention
Example should not be construed as limiting the practical range of the present invention.All skills for utilizing technical solutions according to the invention or this field
Art personnel design similar technical solution and reach above-mentioned technique effect under the inspiration of technical solution of the present invention, or
To all the changes and improvements made by application range etc., the patent that should all still belong to the present invention covers within protection domain.
Claims (1)
1. a kind of woollen and cashmere recognizer based on Gabor wavelet analysis, it is characterised in that：Including online identification process and mould
Type learning process：
The online recognition flow carries out qualitative analysis to the fibre image acquired in real time, includes the following steps：
(1) acquisition of image coordinates Olympus CX41 biomicroscopes, to Cashmere and Woolens using 3,000,000 pixel technical grade ccd
Fiber carries out capture；
(2) it pre-processes：A carries out smothing filtering using Gaussian filter to image, to remove the noise in image, Gaussian filter
It is a kind of low-pass filter, process can be using Formal Representation as the volume of input picture I (x, y) and gaussian kernel function G (x, y)
Product：
S (x, y)=I (x, y) × G (x, y；σ) wherein
B adjusts image gray levels and realizes image enhancement, if data xijIt is the i row j column elements in image X, maxx, minxRespectively
It is maximum, the minimum value in X；
(3) image object extracts：Using the extraction of edge detection and profile based on canny, canny edge detections are more than one
Grade edge detection algorithm；Its basic step mainly has：A obtains x, the gradient of y, and the non-maximum values of b inhibit, c Edge tracks, here directly
It connects using the canny operators inside opencv；
(4) Gabor wavelet feature is extracted：Divide 5 40, the direction of scale 8 Feature Descriptors；Using PCA dimensionality reductions to 100 dimensions；
Gabor functions can extract relevant feature on frequency domain different scale, different directions；Two-dimensional Gabor function can be with table
It is shown as：
Wherein：
The Gabor bases for choosing one group of different scale and direction carry out convolution to input fibre image, specifically choose 4 scale (v=
0,1 ..., 3), 8 directions (i.e. K=8, u=0,1 ..., 7), totally 32 Gabor kernel functions；
(5) feature vector is inputted into classifier calculated result；Each fibre image will be inputted and be uniformly scaled 48*48 sizes,
Then the Gabor characteristic collection of illustrative plates under 32 scales is extracted, extracts 48*48*32=73728 dimensional vectors altogether；Further we take
Original high dimensional data is down to 100 dimensions by PCA methods, the input feature value as SVM；
The model learning flow is to obtain a grader, takes a kind of sorter model based on SVM, including following
Step：
(a) premise of model learning is the database of a large amount of woollen and cashmere of accumulation；
(b) herein on basis, by the way of manually marking, make type and the location of machine hard objectives fiber,
It is a kind of supervised learning mode；
(c) fibre image in database is pre-processed, feature extraction, (2) (4) step of the step and online recognition flow
Unanimously；
(d) learning process takes the SVM classifier of two classification, and is trained using the libSVM to increase income, chooses RBF radial direction bases
Kernel function, iterations are set as 100000 times, iteration ends deviation 0.001.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610091974.4A CN105678341B (en) | 2016-02-19 | 2016-02-19 | A kind of woollen and cashmere recognizer based on Gabor wavelet analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610091974.4A CN105678341B (en) | 2016-02-19 | 2016-02-19 | A kind of woollen and cashmere recognizer based on Gabor wavelet analysis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105678341A CN105678341A (en) | 2016-06-15 |
| CN105678341B true CN105678341B (en) | 2018-11-13 |
Family
ID=56305598
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610091974.4A Active CN105678341B (en) | 2016-02-19 | 2016-02-19 | A kind of woollen and cashmere recognizer based on Gabor wavelet analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105678341B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106503681B (en) * | 2016-10-28 | 2019-05-03 | 东华大学 | A kind of digital picture stage division of wool scale layer |
| CN109948405A (en) * | 2017-12-21 | 2019-06-28 | 中玉金标记（北京）生物技术股份有限公司 | Identification seed direction method based on artificial intelligence |
| CN108036746B (en) * | 2017-12-26 | 2019-08-06 | 太原理工大学 | A kind of Gabor transformation realization carbon fibre composite surface texture analysis method based on Spectrum Method |
| CN108776785A (en) * | 2018-06-01 | 2018-11-09 | 上海工程技术大学 | A kind of woollen and cashmere recognition methods based on multi-feature fusion |
| CN109583308A (en) * | 2018-10-31 | 2019-04-05 | 东华大学 | A kind of Cashmere and Woolens fiber automatic identifying method based on drop shadow curve |
| CN116823830B (en) * | 2023-08-29 | 2024-01-12 | 江苏恒力化纤股份有限公司 | Textile appearance flatness assessment method based on multispectral image |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102495999A (en) * | 2011-11-14 | 2012-06-13 | 深圳市奔凯安全技术有限公司 | Face recognition method |
| CN102663426A (en) * | 2012-03-29 | 2012-09-12 | 东南大学 | Face identification method based on wavelet multi-scale analysis and local binary pattern |
| CN103605958A (en) * | 2013-11-12 | 2014-02-26 | 北京工业大学 | Living body human face detection method based on gray scale symbiosis matrixes and wavelet analysis |
| CN103824057A (en) * | 2014-02-18 | 2014-05-28 | 江苏大学 | Pig respiratory rate detection method based on area feature operator |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8798374B2 (en) * | 2008-08-26 | 2014-08-05 | The Regents Of The University Of California | Automated facial action coding system |
| US8917914B2 (en) * | 2011-04-05 | 2014-12-23 | Alcorn State University | Face recognition system and method using face pattern words and face pattern bytes |
| TWI469087B (en) * | 2011-04-26 | 2015-01-11 | Univ Nat Cheng Kung | Method for depth map generation |
| KR101877981B1 (en) * | 2011-12-21 | 2018-07-12 | 한국전자통신연구원 | System for recognizing disguised face using gabor feature and svm classifier and method thereof |
-
2016
- 2016-02-19 CN CN201610091974.4A patent/CN105678341B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102495999A (en) * | 2011-11-14 | 2012-06-13 | 深圳市奔凯安全技术有限公司 | Face recognition method |
| CN102663426A (en) * | 2012-03-29 | 2012-09-12 | 东南大学 | Face identification method based on wavelet multi-scale analysis and local binary pattern |
| CN103605958A (en) * | 2013-11-12 | 2014-02-26 | 北京工业大学 | Living body human face detection method based on gray scale symbiosis matrixes and wavelet analysis |
| CN103824057A (en) * | 2014-02-18 | 2014-05-28 | 江苏大学 | Pig respiratory rate detection method based on area feature operator |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105678341A (en) | 2016-06-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105678341B (en) | A kind of woollen and cashmere recognizer based on Gabor wavelet analysis | |
| CN110837768B (en) | Online detection and identification method for rare animal protection | |
| CN105913081B (en) | SAR image classification method based on improved PCAnet | |
| Ibrahim et al. | Leaf recognition using texture features for herbal plant identification | |
| Zawbaa et al. | An automatic flower classification approach using machine learning algorithms | |
| CN104021384B (en) | A kind of face identification method and device | |
| Harish et al. | Classification of plant leaves using Morphological features and Zernike moments | |
| CN107292314A (en) | A kind of lepidopterous insects species automatic identification method based on CNN | |
| CN103345617A (en) | Method and system for recognizing traditional Chinese medicine | |
| Rahamathunnisa et al. | Vegetable disease detection using k-means clustering and svm | |
| Revathi et al. | Homogenous segmentation based edge detection techniques for proficient identification of the cotton leaf spot diseases | |
| Amlekar et al. | Leaf features based plant classification using artificial neural network | |
| Ouyang et al. | The research of the strawberry disease identification based on image processing and pattern recognition | |
| Jana et al. | Intra-class recognition of fruits using color and texture features with neural classifiers | |
| CN104268587B (en) | False fingerprint detection method based on finger wave conversion and SVM | |
| Pushpa et al. | Comparision and classification of medicinal plant leaf based on texture feature | |
| Talasila et al. | PLRSNet: a semantic segmentation network for segmenting plant leaf region under complex background | |
| CN108710836A (en) | A kind of lip detecting and read method based on cascade nature extraction | |
| Haarika et al. | Insect classification framework based on a novel fusion of high-level and shallow features | |
| Sethy et al. | Pest Detection and Recognition in Rice Crop Using SVM in Approach of Bag-Of-Words | |
| Patil et al. | Identification of growth rate of plant based on leaf features using digital image processing techniques | |
| Fu et al. | Recognition of plants with complicated background by leaf features | |
| Hong-Ying et al. | An iris recognition method based on multi-orientation features and Non-symmetrical SVM | |
| Handa et al. | A review and a comparative study of various plant recognition and classification techniques using leaf images | |
| Ghosh et al. | A leaf image-based automated disease detection model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C41 | Transfer of patent application or patent right or utility model | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20161011 Address after: 300193 Tianjin city Nankai District Road Bridge No. 25 Applicant after: TTTS DETECTION TECHNOLOGY CO., LTD. Applicant after: Kaisedaike Environmental Protection Tech Co., Ltd., Tianjin City Address before: 300000 Tianjin city Nankai District Road Bridge No. 25 Applicant before: TTTS DETECTION TECHNOLOGY CO., LTD. |
|
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: 300193 Tianjin city Nankai District Road Bridge No. 25 Applicant after: Tianfang standard testing and certification of Limited by Share Ltd Applicant after: Kaisedaike Environmental Protection Tech Co., Ltd., Tianjin City Address before: 300193 Tianjin city Nankai District Road Bridge No. 25 Applicant before: TTTS DETECTION TECHNOLOGY CO., LTD. Applicant before: Kaisedaike Environmental Protection Tech Co., Ltd., Tianjin City |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20211202 Address after: No. 158 hanghai Road, pilot free trade zone (Airport Economic Zone), Binhai New Area, Tianjin 300308 Patentee after: TIANFANG STANDARD TESTING AND CERTIFICATION CO.,LTD. Patentee after: Tianjin Xieli Automation Engineering Co., Ltd Address before: No. 25 Magpie Bridge Road, Nankai District, Tianjin Patentee before: TIANFANG STANDARD TESTING AND CERTIFICATION CO.,LTD. Patentee before: Tianjin Kaiser dyke Environmental Protection Technology Co., Ltd |