Cloud Detection Method for High Resolution Remote Sensing Imagery Based on the Spectrum and Texture of Superpixels.

Image cloud detection is an important part of high spatial resolution remote sensing imagery (HSRI) information processing and analysis. The spectral threshold selection of image cloud detection and the influence of cloud-like ground objects are two vital factors in determining cloud detection results of HSRI. With respect to these two issues, a novel cloud detection method for HSRI based on the spectrum and texture of superpixels is proposed in the paper. First, the adaptive image cloud detection spectral threshold is obtained according to the image equalization histogram. Second, the initial cloud detection result is obtained based on spectral threshold of the cloud detection and spectral attributes of superpixels. Third, the initial cloud detection result is refined based on the gray value and angular second moment of the superpixels local binary patterns texture to eliminate the influence of cloud-like ground objects. Finally, the cloud detection result is processed using the region growing algorithm and expansion algorithm to obtain an accurate cloud detection result. The experimental results show that the proposed method can obtain good cloud detection results. [ABSTRACT FROM AUTHOR]