基于支持向量机的遥感影像厚云及云阴影去除

本文提出了一种基于支持向量机的遥感影像厚云及云阴影去除方法。首先利用支持向量机的学习性能检测影像中的云层，并利用太阳角度信息，判定云阴影区域，得到云层和云阴影的二值图。再对影像进行支持向量值轮廓波变换，利用云层和云阴影二值图生成的选择矩阵，对变换系数进行多层镶嵌，完成云层及云阴影的初去除。对影像镶嵌未能去除的云层及云阴影，通过统计学补偿的方法进行修复。最后重构图像并进行中值滤波实现厚云及云阴影去除。仿真实验表明，该方法能更好地再现云层覆盖区域的地物信息，去云后的图像具有更好的光滑度和清晰度。

The Removal of Thick Cloud and Cloud Shadow of Remote Sensing Image Based on Support Vector Machine

An approach of removing thick cloud and cloud shadow in remote sensing image based on support vector machine is suggested.Firstly,the learning ability of support vector machine is used to detect cloud in remote sensing image,and combining the information of solar angle,cloud shadow area is detected.So the binary images of cloud and its shadow are obtained.Secondly,the remote sensing images are transformed by support vector value contourlet transform.The transforming coefficients are mosaiced using selection matrices produced by the binary images to achieve preliminary removal of cloud and its shadow.Finally,the cloud and its shadow which can not be removed by image mosaic are repaired by using the method of statistics.Experiments show that the method can recover the ground information covered by cloud efficiently and reconstruct the cloud free image with clear details and better smoothness.