Study On The Foreign Matter Recognition In Cotton Using Image Processing

The amount of cotton textile production is very large in our country. But the foreign matters and impurities in cotton seriously influence the cotton production and quality which result in great loss in both economy and non-economy. The research and design of foreign matter recognition and elimination in cotton is precisely under this strong application background. Considering the large amount of the cotton textile production, those machine equipments overseas are too expensive to be afforded. Therefore the research and exploitation independently on foreign matter recognition and elimination, to find the most needed and relative lower price and the superior performance system, has the profound significance and influence.The recognition algorithm and elimination method, which are based on digital image processing technology and the signal control technology are the keys of this paper. On the basis of thorough analysis to the foreign matter image, this paper put forward several recognition algorithms, such as edge detection on image mold filter, improved Ostu thresholding, image recognition on fractal-based, adaptive threshloding based on fractal dimension and so on. This mathematical model and algorithm proved strictly in mathematics have compared with each other objectively which is advantageous to realize the standardization in the promoted application. The result based on a series of contrasted image experiments shows that the method of adaptive threshloding based on fractal dimension is good for the foreign matter recognition. For the most difficult recognizing foreign matters such as the white polypropylene fibers, the way of recognition on fractal-based taking the texture as parameter has been proposed in particular. The fractal method has given a new mentality for foreign matter recognition in cotton. The recognition algorithm, applying skills and the research experience introduced by this paper can be also transplanted and applied to other projects in textile domain.This full paper with six chapters has discussed several algorithms of foreign matters recognition in cotton based on image techniques. In the first chapter the background and significance of this project are narrated and the development of relative studies in this filed are also introduced. In the second chapter the whole system design frame and hardware platform built are broadly introduced. In the third chapter several recognition algorithms regarding cotton images are. In the fourth chapter the design and control of elimination parts are introduced with emphasis. In the chapter followed the effect of image algorithms mentioned before are evaluated through a series of experiments. In the last chapter the result and the further works to do are put forward.