MORPHOLOGICAL EXTRACTION OF LINE NETWORKS FROM NOISY LOW-CONTRAST IMAGES

Novel algorithms for line and line network extraction from images are proposed. The algorithms which operate on region adjacency graphs are developed especially for images that are noisy and have low contrast. Microscopic images of fibers are used as an example application. The p roposed algorithms are very general in nature, since they can detect l ines of arbitrary geometry. The region adjacency graph is constructed from a segmented version of the original image. The segmentation proce ss is based on the watershed transformation which is widely used in mo rphological image analysis. A Laplacian-type operation for line detect ion on the region adjacency graph is presented and removal of noise st ructures is studied. Fast algorithms for the detection of line ends an d line branches are proposed. In addition, a directional propagation a lgorithm is derived. The line-end and branch-detection algorithms as w ell as directional propagations use only the adjacency information of the region adjacency graph. Finally, a complete line-network-extractio n system is presented. The performance of the proposed algorithms is s tudied and a comparison between algorithms operating on the square gri d and on the region adjacency graph is made. (C) 1995 Academic Press, Inc.