Clouds and their shadows in remote sensing images bring many troubles to image interpretation and application.It is important to detect and repair these regions rapidly.An efficient method was proposed to detect and repair the unwanted clouds and shadows using two images of the same site acquired in adjacent time.At first

the cloud and shadow regions were enhanced by fusing multi-spectral data and then detected automatically using Otsu’s algorithm.Since the clouds and their shadows will lead to obvious differences in brightness between the two images

false detection of highlighted objects and water regions can be easily discarded

and then the normalized cloud and shadow density map can be built.Linear combination of two images and histogram matching algorithm were adopted to repair the shadow and cloud regions.Experiments on SPOT 4 multi-spectral remote sensing data show impressive results which will benefit several applications.