Research On Adhesion Of Grain Particles Image Segementation

With the rapid development of information technology, image processing technology has become one of the important tools that are in the fields of the current scientific research and industry, agriculture, medicine and others.Grain particles image reflects the shape of grain particles, and their large amount of information processing more difficult. Grain particles image segmentation is a key part of the research on grain particles image, testing for food quality has important application value.There are many methods of grain image segmentation, according to the morphological characteristics of grain particles. In this paper, it is researched on mathematical morphology and segmentation of watershed transformation, improved morphological gradient amendment algorithm, verified by experiment. The following is the main work of this paper.This paper improves gradient image algorithm of grain particles based on morphological gradient transformation. In this paper, for the irregularity of grain particles, and it is difficult to select morphological structure element, the method of selecting multi-scale morphological structure element is introduced that is fit to segment grain particles image, the gradient algorithm that computes difference between the multi-scales morphological gradient mean is proposed.This paper focuses on the algorithm of morphological amendment for grain particles image. With the purpose of filtering out the noise of gradient image, using features which morphological amendment algorithm can smooth gradient edge, an amendment method that morphological amendment gradient mean is presented. The algorithm could smooth the irregular detail points and small noise of gradient image, and can enhance the contrast of gradient image. To further smooth the smaller noise in contours of marker image, a new amendment method is comprised by combining the distance transform with the opening and closing amendment to smooth the edge of gradient marker image. The method can well remove the smaller noise in the marker, and has amended the marked outline. Finally, the image is used watershed algorithm to segment.In this paper, a method of improved and based on morphological watershed segment is given. Compared with the traditional watershed algorithm and the exiting researched result, the algorithm of this paper can better overcome the over-segmentation in watershed segmentation algorithm, is more suitable to segment adhesive grain particles image.