Feature Extraction Method Researches of SAR Targets Based on 2-D Wavelet Transform

摘要

Edge detection is the basis of image analysis. The influence of speckle within SAR images should be taken into consideration during edge extraction. The traditional methods of edge detection based on space domain, such as Sobel, Roberts and Laplacian operator, are ineffective due to poor de-noise capability. The 2-D wavelet transform is introduced in this paper for its satisfactory local characteristics. The effect of speckle on edge extraction is suppressed by adopting threshold. Experiment results clearly show this method is efficient.