Contrast Enhancement Technique for Efficient Detection of Cloud from Remote Sensing Images

Satellite imaging is essential for various applications, including disaster management and recovery, agriculture, and military intelligence. Clouds are a severe impediment to all of these applications, and they must be customarily identified and removed from a dataset before satellite images can be used for further processing. The quality of the satellite images is affected by various factors during the acquisition process. In this paper, an enhancement approach is proposed to improve the quality of the satellite images to improve the accuracy of the cloud detection process. The enhancement process utilizes the edge information extracted from the input image. The extracted edge information creates a variational map to equalize the intensities by distributing them to occupy the whole dynamic gray scale. Experiments have been performed to validate the efficiency of the enhancement process on the segmented results. The analysis shows that the enhancement process aids in improving the cloud detection, which is indicated by the high values of the performance measures such as accuracy, F1-score, Dice, and Jaccard coefficient compared with the un-processed images from the sentinel-2 remote sensing dataset.