论文部分内容阅读
提出了一种有效针对多光谱遥感影像的云影检测与阴影区域修复方法。基于同一地区时相相近的两幅影像,充分利用碎云及阴影的光谱特性分别对云影区域进行融合增强,然后采用Otsu算法求解最佳阈值自动检测出云及阴影区域,根据云影的出现会引起两幅影像局部相应区域明显的亮度变化,可排除亮地物和水体的影响,建立归一化的云影密度图,在此基础上,采用线性加权组合与光谱直方图匹配相结合的方法对其加以修复,利用SPOT4影像进行的实验表明其修复效果完全能够满足应用需要。
A new method of cloud shadow detection and shadow area restoration for multispectral remote sensing images is proposed. Based on two images with similar phases in the same region, the cloud-shadow region was fused and enhanced by using the spectral characteristics of cloud and shadow respectively. Then, the optimal threshold was calculated by Otsu algorithm to automatically detect the cloud and shadow regions. It will cause obvious brightness changes in the corresponding parts of the two images and eliminate the influence of the bright objects and the water body to establish a normalized cloud density map. Based on this, a combination of linear weighted combination and spectral histogram matching Methods to be repaired, the use of SPOT4 images of the experiment shows that the repair effect can fully meet the application needs.