Fabric Detect Detection And Classification Method Based On Machiine Vision

China is a country with a large production of textile,and the textile industry in our country has been occupies a pivotal position in the national economy.However fabric defects point to impact the quality of products,not only to textile enterprises economic interest loss,also seriously affected our country textile in the international competition.Using machine vision technology to flaw detection and classification of the textile products to improve the quality of textile products has the important value of application research.To automatic fabric defect detection and classification the most critical part is the defect segmentation,defect marking,and defect feature extraction,the result directly affect the accuracy of detection and recognition.To solve these problems,this paper carried out relevant research and analysis,mainly done the following research work:(1)In order to prevent the acquisition of image is too dark or too bright,using gray contrast enhancement on the sample image gray stretch.according to the characteristics of sample in the image noise is gaussian distribution,usintg the gaussian filtering on the sample image denoising processing.(2)In terms of fabric defect image segmentation,first has carried on the uniform to image block processing,then analyzes the different methods of binarization of defect segmentation effect,finally,by adopting the method of OTSU as defect segmentation algorithms,a better segmentation results have been achieved.(3)For miscellaneous points in the image after segmentation,using morphological processing and connected component analysis method to remove the noise points,implements the markup of fabric defect.(4)According to to the local characteristics of fabric defect,this paper studies the SIFT feature extraction method,and put forward a kind of classification method based on SIFT features,K-means clustering and the LDA theme model combining,with several common types of defects can be better classification.(5)Under the environment of Microsoft Visual Studio 2010 compilation,combining OpenCV image processing library,to develop the fabric defect detection and classification system based on machine vision,through the experimental analysis,further verify the algorithm in detection,marking,and classification of defects is effective.