Research Of Segmentation Quasi-circular Region Based On Gabor Mmethod

The circular objects are previous in industrial production and real life, therefore the segmentation of circular object form image plays an important role in a great deal of field. The automatic account process which is established byJiangsu Shagang Group and csu is good solution for problem of the automatic bar production. In order to achieve system performance without supervision, comprehensively improve the level of automation systems, the study of unsupervised segmentation for bundle of rod should be implemented. In this paper, the novel framework which can be used in the circular object segmentation (like bars, pills, corn and green) is proposed for this issue.First, the texture image of circular objects which is based on frequency domain analysis from a new way is found and analyzed. Comparing all kinds of the methods frequency domain analysis and studying the drawback of common methods, the way of unsupervised Segmentation for the area which is full with the circular object is established.Secondly, The selection of Filter is based on dispersion of response, this way improve the efficient for characteristic response. Considering the common ways reduce the response information farther, the algorithm of bridge fuzzy is proposed. This method can maximum the data form response result filter, and then improve the efficient of segmentation.Final, flow chart based on the new advanced Gabor algorithm for circular object segmentation is proposed. In contrary on the common ways, this framework have obtained the good results in experiment...