一种陆地卫星影像厚云阴影检测方法

Detection of thick clouds shadow from Landsat TM data

针对现有的遥感影像厚云阴影监测方法存在过程复杂、精度不高等问题,该文提出一种TM影像厚云阴影检测方法:通过阈值分割提取影像中的所有阴影;对比一幅近时相的无云影像,分离出厚云的阴影,即云阴影为"有云"影像比"无云"影像多出的阴影;最后将提取结果与人机交互、云阴影增强模型法提取的结果进行对比分析。实验结果表明该方法提取精度较高,可以快速、准确地提取TM影像中的厚云阴影。

Aiming at the problem that current thick clouds monitoring methods of RS imagery are complex and low-accurate,the paper presented a method based on TM imagery：all the shadow in the TM image was extracted with threshold segmentation;a cloud-free image of recent phase was compared with the cloud image to separate the shadow of clouds,thus clouds shadow was got as the part where the shadow is more;finally,the extracted result was comparatively analyzed with that of the human-computer interaction methods and clouds shadow enhancement model respectively.Result showed that the method could extract the thick clouds shadow from TM imagery efficiently with high accuracy.