Study on cloud processing methods with MODIS data

Owing to the influence of climate, cloud noise is one of the most disturbance factors in remote sensing image. Therefore, cloud detection and removal has become a very important step before using MODIS data. In this paper, the multispectral cloud detection algorithm based on spectral character and threshold is used to detect cloud, then, different methods to remove the cloud be used according to different cloud thickness in images. At the same time, in order to compensate for the defects of different removing cloud methods, the improved homomorphism filtering algorithm and replacement method are selected to remove cloud in MODIS images. Finally, it is found that the information of cloud-free is not changed and the effect of cloud removal is very obvious.