Improved Chan-Vese Model and Its Application on High Resolution Remote Sensing Image Segmentation

Abstract:

An improved CV model was presented according to the characters and the segmentation requirements of the remote sensing images. The correctness of the improved model was validated by the experiment. The results show that the new method can improve calculation efficiency effectively. Beside this, a continuous and closed boundary curve of the target objects can be obtained at the same time. Therefore, the land cover can be identified precisely with the help of the geometrical characteristic of the segmentation boundary.