Traditional Chinese Patterns Analysis Based on Moment Invariants

摘要

In this paper, the numerical analysis is studied with the help of computer technology to geometry texture features of traditional Chinese art patterns. First, the gray-scale distribution of an original image is adjusted by using the gray value transformation and a new image with five gray levels is formed. Second, by using the geometry invariant moment method in image analysis and processing, 12 geometry invariant moments of the new gray image as its geometric texture feature vector is calculated. Lastly, it carries on the classification using k-means clustering method in clustering analysis. Through specific examples, it shows that the method is simple in principle, easy realized, and it will provide a reference method for computer analysis and the processing of traditional Chinese art pattern.