The Recognition of Linear Objects Based on Edge Detection and Skeleton Tracking

Abstract

The recognition of linear objects is applied to many aspects, but most of usual image recognition schemes need to be improved in terms of accuracy, anti-interference ability etc. This paper presents a method to recognize linear objects that synthesizes the edge and the skeleton of linear object. Directional wavelet transform not only presents the multi-scale properties, but also can better reflect the edge information of images, because direction and texture information of the image can be better expressed. The skeletons tracking method based on direction priority for searching can decrease searching region and reduce the complexity. This paper presents a robust method to recognize linear object that synthesizes the edge extraction based on directional wavelet transform and the skeleton tracking based on direction priority for searching. Our experiments suggest that this method is robust and reliable in dealing with pulp fiber images.