Remote sensing imaging simulation and cloud removal

Abstract

Cloud obstacles obscure ground information frequently during remote sensing imaging which leads to valuable information losses. Removing clouds from a single image becomes challenging since no reference images containing cloud-free regions are available. In order to study cloud removal technologies and evaluate their performances, a method to simulate evenly and unevenly distributed clouds was proposed by analyzing the physical model of remote sensing imaging. Dual tree complex wavelet transform (DTCWT) and its features were introduced briefly. According to the frequency relationships between clouds and ground objects in remote sensing images, a novel cloud removal algorithm was proposed. The algorithm divided the cloud-contaminated image into low-level high frequency sub-bands, high-level high frequency sub-bands and low frequency sub-band by DTCWT. Low-level high frequency sub-bands were filtered to enhance the ground object information by Laplacian sharpening. The other two types of sub-bands were processed to remove clouds by cloud cover coefficient weighting (CCCW). The experiments were implemented to process cloud disturbed images produced by the proposed simulation method. The results of cloud removal from remote sensing images were analyzed. It proved the proposed algorithm is greatly superior to algorithms based on traditional wavelet transform and dark channel prior.