A Water/Land Segmentation Algorithm Based on an Improved Chan-Vese Model with Edge Constraints of Complex Wavelet Domain

A novel model for segmentation of water/land in Synthetic aperture radar (SAR) images is presented. The proposed model calculates the global matrix of oriented gradient of histograms on wavelet subbands based on the multi-scale property, shift invariance and directional selectivity of Complex wavelet transform (CWT), then acquires edge information by thresholding the matrix. It improves the traditional Chan-vese (CV) model by adding the constraint of extracted edge information. A new water/land segmentation algorithm is then proposed based on the new model. The performance of proposed method is compared with some other existing algorithms and the experimental results confirm its effectiveness.