A New Space Image Segmentation Method Based on the Non-Subsampled Contourlet Transform

Abstract

Image segmentation is very essential and critical to image processing and pattern recognition. In order to extract the feature stars, a novel method for image segmentation by use of nonsubsampled contourlet transform (NSCT) is provided by this paper. The space images are transformed by the NSCT, and then the magnitude of the NSCT coefficients of all the directional subbands at a specific level are computed and the magnitude are compared with the adaptive multilevel threshold which is gained by the Lorentz information measure (LIM), so the edges of the stars are extracted. Using the measure of consistency (F) as the evaluation criterion and compared with the Otsu method, the preliminary experimental results demonstrate the robustness and efficiency of the proposed algorithm in the presence of noise in the space image and the high accurate segmentation of this algorithm.