The Research Of Skeleton Representation And Similarity Measurement For Profiled Fiber Images

With the emergence of shaped fiber, Fibers blended fabrics are more and more be favored by the people in the international market. But the style, performance and price of fabrics are affected by the content of fiber fabric. So it is very important to recognize the ingredient of fiber fabrics. Due to inaccuracy and high time complexity of manual fiber recognition, the computer aided fiber recognition has gotten more and more attention. However, automatic fiber recognition is still a relatively complicated research and no research result has been applied successfully. This paper is a part of the research, which is sponsored by the Foundation of National Excellent Doctoral Dissertation of P. R. of China and the Scientific Research Foundation for the Returned Overseas Chinese Scholars, State Education Ministry, and is on the specific demands of Shanghai Entry-Exit Inspection and Quarantine Bureau on textile and fiber test. The automatic recognition of the natural fiber as well as shaped fiber is the key of the project. This paper is a part of the automatic recognition of the natural fiber and research the method of feature extraction and recognition of Shaped fibers.As the Shaped fibers were microscopic amplification, which exist some warping. So it is have no steady and high fibers recognition rate by the commonly method which based on mathematical morphology. Also because the shaped fiber skeleton structure is simple and different fiber cytoskeleton is based on similarity. So it is have no achieve exact fibers recognition by the commonly method which based on the skeleton of single feature. So a hierarchical and skeleton-based algorithm which also combines with other geometrical features was proposed. This method has the advantage of skeleton description, which can solve the problem of warping of shaped fibers. Also the skeleton shape and fiber outline of geometric features can be got, which can solve the problem of the simple and similarity of fiber skeleton. Skeleton as complete and accurate shape descriptors, with the original graphic of the same topological structure, can retain the original the shape of the information, has the good translation, rotation and scale variations of invariant. Therefore in this paper, the skeleton information was mapped to a skeleton tree structure by contour tracking algorithm, then a skeleton tree's adjacency matrix was constructed and eigenvalues of the matrix was worked out, after that, the profiled fiber image was divided into several categories according to the eigenvalues.Due to the shaped fiber skeleton is simpler and the skeleton have similarity, so it is only can divided the shaped fibers into several categories or not reached a classification of precise identification, which just rely on shaped fiber skeleton matrix eigenvalues. In this paper, in order to solve this problem, the shaped fiber images are continues to differentiate after which are divided. According to every kind of fiber image skeleton characteristics and contour geometric feature, the skeleton perimeter statistical characteristics and fiber contour concave dot characteristics were gotten. The experimental results show that the profiled fiber image can be accurately recognized by the proposed method.