| China is a big country in textile ,The amount of cotton textile production is very large inour country. But the foreign matters and impurities in cotton seriously influence the cottonproduction and quality which result in great loss in both economy and non-economy. Theresearch and design of foreign matter recognition and elimination in cotton is precisely underthis strong application background.The recognition rate of foreign body in raw cottondetermines the rate of the size of the removed .How to identify the presence of raw cotton invarious foreign bodies quickly and accurately, thereby enhancing the rate of foreign bodyremoved, it is the content and purpose of this study.Paper identify foreign objects on raw cotton based on MATLAB image processingtechnology.On the basis of thorough analysis to the foreign matter image, this paper putforward two recognition algorithms, such as improved Ostu thresholding, adaptivethreshloding based on fractal dimension . Papers selected representative images of threeforeign raw cotton. After pretreatment in these three image, Using improved Ostuthresholding method and the fractal dimension based on adaptive threshold methodrespectively,Using a series of mathematical morphological approach to identification of rawcotton in the foreign body, identifing the foreign body from The raw cotton.Analysising and comparisoning the two threshold based segmentation of raw cottonforeign body recognition effect. We know that ,when the color difference is large betweenraw cotton and foreign body , it can identify all of the foreign body with the two thresholdmethod, and the results are good; when the color of raw cotton and foreign bodies is similar,the latter method is better than the former method in the effect of identification, but the lattermethod need more time and Labor force than the former one.Therefore, the methods used to identify foreign objects on the raw cotton should be based onthe actual circumstances. |
PDF Full Text Request