Research On Fast Processing Method For Color Images Of Foreign Fibers In Cotton

Aimed at online recognition of foreign fibers in lint, a fast processing method for foreign fibers in cotton was proposed. Automated visual inspection and image processing are main techniques used in this research. The main research works are as follows:(1) Edge detection method for color images of foreign fibers in cotton. A new edge detection method was developed to detect the foreign fiber objects out of the background. In this method, the characteristics of RGB in color images were analyzed, and the edge of color images were detected by an edge detection method based on improved mathematical morphology. The color images were subsequently converted into gradient maps. The experiment results indicate that the proposed method successfully detects the edge of high-resolution color images for foreign fibers in cotton both directly and precisely. This operator can also solve the problem that conventional edge detection operators are not accurate enough for segmentation of high-resolution images of foreign fibers.(2) The determination method of best threshold for foreign fibers in cotton. An improved iteration threshold method was used to segment the gray gradient maps. The relationship and law between final threshold and previous threshold are analyzed by using the iteration threshold method in segmenting images. And then the empirical value, which is fit for the gray gradient maps of most foreign fibers, is determined. At last 7×7 two-dimensional filtering method was employed to remove away the small objects generated by noises and pseudo-foreign-fibers. Thus the clear binary image is obtained. The segmentation results indicate that the method presented in this paper can achieve the accuracy for image segmentation of foreign-fiber, and the speed of image processing is much faster than the conventional methods.(3) Design and implementation of the image processing system of foreign fibers in cotton. The objective of the image processing system is to test the above mentioned methods. The hardware system of the image processing includes 6 parts, i.e., lint input conveyer, lint layer generator, output roller, image acquisition system, lint layer collecter and host computer. The software system includes 4 modules, i.e., parameter setting, image acquisition, image segmenting, small objects removing. Great deals of experiments indicate that the developed method, a fast processing method for color images of foreign fibers in cotton, works effectively and efficiently.