改进CV模型在高分辨率遥感影像分割中的应用

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

针对遥感图像的特点及分割要求给出了一种CV简化模型,并对改进模型的正确性进行了实验验证。实验结果表明,该方法不但提高了运算速度,而且能够得到连续封闭的目标地物矢量数据,因此可方便地利用分割边界的几何特征实现地物目标的精确识别。

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.