Abstract: To extract abundant edge information and improve the ability to resist noise, a noisy image edge detection algorithm is proposed based on Non-subsampled Contourlet Transform. The proposed approach adopts a hard threshold method to remove the noise. An adaptive threshold approach is used to obtain the high frequency edge information. The low frequency edge image is achieved according to the low frequency matrix and is adopted to compensated the high frequency image for obtaining the final edge image. Experiment results show the effectiveness and excellence on different noisy images of the proposed algorithm.

Key words: edge detection, NSCT, noisy image, hard threshold method