遥感影像云团和云影检测与去除

Detection and Removal of Clouds and Their Shadows from Remote Sensing Images

以图像处理和分析理论为依据,建立了云团、云影图像模型和区域恢复模型,并采用阈值法对图像分割。用1988年TM遥感影像对2002年ETM遥感影像中云团及云影进行去除试验,再现云团和云影区地物特征,提高遥感影像数据的可用性。结果表明,使用阈值法并利用直方图形态分析选取阈值、检测云团和云影的计算方法简单且效果好;选取合适的两遥感影像灰度平均值之比,可以得到理想的恢复效果,再现云团和云影区地物特征,达到提高遥感影像的可用性。

This paper studies the detection and removal of clouds and their shadows from remote sensing images based on the theory of image processing and analyzing, and sets up the image model and region restoring model. The image segment is made by the method of thresholding. Tests on removal of clouds and their shadows from ETM remote sensing images of the year 2002 with TM remote sensing images of the year 1988 represents the features of objectives on the Earth＇s surface in clouds and their shadow region. The results show that, it is found that the computation method is simple and effective by adopting threshold method that the threshold value is chosen according to the shape of histogram. Selecting proper gray scale ratio of twin remote sensing images may obtain ideal restoring results, which may represent the objectives on the Earth＇s surface in the clouds and shadow, and upgrade the application od remote sensing images.