彩色遥感图像薄云去除方法

针对彩色遥感图像中薄云带来的降质问题,提出了一种基于低秩矩阵分解的去云方法。将彩色遥感图像分成三个单通道图像,对每一个通道进行低秩矩阵分解,得到单通道薄云信息;根据薄云在三个通道中均匀分布的特点,选取合适的阈值对薄云信息进行自动判定,提取代表三个通道的薄云图像;用三个通道的图像分别减去对应通道的薄云图像,并融合三个通道的结果得到去云后的彩色遥感图像。实验结果表明,该方法不仅能够保留无云区域信息的完整性,而且对有云区域的处理在主观视觉效果和客观评价指标上都具有较好的效果。

Thin cloud removal method in color remote sensing image

A thin cloud removal method via low-rank matrix decomposition is proposed to solve the problem of reduc-ing the image quality of thin cloud in color remote sensing image .A color remote sensing image is divided into three single channels ,and each channel is decomposed into low rank matrix to obtain the single channel thin cloud information .Ac-cording to the characteristics of thin cloud's even distribution in three channels ,the appropriate threshold is selected to determine the thin cloud information automatically ,and the thin cloud images representing the three channels are extrac-ted .The cloudless color remote sensing image is obtained by fusing three channel images minus the corresponding channel cloud images respectively .The experimental results show that the method not only can retain the integrity of information in the cloudless regions ,but also has a good effect on the subjective visual effect and the objective evaluation index in pro-cessing the cloud regions .