Cloud removal using efficient cloud detection and removal algorithm for high-resolution satellite imagery

Abstract

This paper describes a technique - efficient cloud detection and removal ECDR algorithm based on remote sensing information. It is obvious that cloud masking has been a challenging risk in receiving the information from the satellite sensor images. For instance, in-paint and multitemporal averaging method, which are existing approaches implemented against the cloud masking but still the requirement has been found inadequate as it has partially contributed to its removal. Hence, the proposed ECDR algorithm paves the way to find accurate information. The cloud-obscured area reconstructed by the ECDR algorithm involves two steps: detecting the cloudy region by pixels values and reconstruction of cloudy regions. For cloud masking, the algorithm calculates the percentage of variance in earlier remote sensing information and, based on the percentage, the cloud pixel value is replaced. Thus, the found ECDR algorithm accuracy is much better than the existing algorithms for deducing information.