Skip to Main Content
SAR (synthetic aperture radar) images are strongly disturbed by speckle noises, which brings difficulty to the segmentation of the SAR images. Some model-based SAR image segmentation approaches (such as the Markov random field model, fractal model) are too complex and time consuming. We propose an new method to segment the SAR images fast in this paper. First we transform the original SAR intensity image into some subimages using the wavelet packets frame transformation. Then we analyze the energy distribution in the transformed subimages and find that the energy of the speckle noise is mainly concentrated in the high frequency subimages. So we segment the SAR image based on the gray level of the low frequency subimages using a threshold classifying method. The result shows that this method can get a satisfying segmentation result and time saving.