Cloud Removal From Multi-temporal Satellite Images Using Information Cloning and Information Reconstruction

In this paper, we  propose a cloud detection and removal approach based on information cloning. The approach removes cloud-contaminated portions of a satellite image and then reconstructs the information of missing data utilizing temporal correlation of multi temporal images[1]. The basic idea is to clone information from cloud-free patches to their corresponding cloud-contaminated patches under the assumption that land cover change insignificantly over a short period of time. Firstly, cloud is detected by using simple thresholding approach. Then a semi automatic approach is used to detect the cloud regions and the SSIM index for both the target and reference images is calculated to sort out according to image similarity. Finally the patch-based information reconstruction is mathematically formulated as a Poisson equation and solved using a global optimization process. Thus, the proposed approach can potentially yield better results in terms of radiometric accuracy and consistency.

Keywords: Cloud removal, information cloning, Poisson equation.