Research On Granular Computing And Its Application In Image Classification

Nowadays, information granule and granular computing are hot studies in the softcomputing. As one of the new methods to manage information and knowledge, granularcomputing has been paid dosed attention by many home and abroad scholars in recentyears, and it has been applied in lots of fields.It's the basic thought of granular computing to compute with information granulewhich is significant in problem description and solution. Granular computing is aproblem describing and solving method, which depends on the abstraction of the realityby granules, the relations among granules, the composition and decomposition ofgranules, and the transformation among granules.On the basis of studying fundamental principles and techniques of granularcomputing, the concerning knowledge is applied in image pattern recognition in thepaper. Against the background of image, the paper stresses on the construction,management and computing of character information granules. What's more, a classifierbased on image character information granular is advanced in the paper by analyzingand studying the image character information granular structure of differentgranularities of characteristic information granule. Besides, simulation classificationtests are done towards six different image classes. It's proved that the classifier based ongranular computing is feasible.Finally, a conclusion is had in the end of the thesis, and analyzes the aspects whichneed to be improved. It also points the future research direction in the imageclassification domain of information granule and granular computing.