A Review of Cloud Detection and Thick Cloud Removal in Medium Resolution Remote Sensing Images

CHEN Jinle,ZHANG Jinshui,DUAN Yaming,YANG Zhi,XIA Xingsheng
DOI: https://doi.org/10.11873/j.issn.1004-0323.2023.1.0143
2023-01-01
Abstract:Medium resolution remote sensing image(10 m~100 m resolution) has moderate spatial resolution, high revisit period and large width to achieve large-scale earth observation, and is currently the core remote sensing data source for accurate acquisition of earth surface information. Studies have shown that more than 60% of the earth’s surface is covered by cloud all year round, which is one of the biggest limiting factors for obtaining effective surface information from medium resolution optical images. How to efficiently tag cloud/shadow and the synthesis of clear image, is the realization of the surface elements extraction, the dynamic change of land cover, and the earth system material and energy circulation parameter inversion of the key, can be regarded as the same as the radiation correction, geometric correction for remote sensing image preprocessing step, it is also a basis for a variety of quantitative remote sensing application. A summary of past studies on cloud detection and thick cloud removal from medium resolution images shows that there have been several review articles on cloud detection up to 2019, but no review articles on thick cloud removal have been reported. Therefore, this paper focuses on summarizing the scientific research results of cloud detection methods since 2018, especially the technical methods based on machine learning, and sorts out the current status and focus of this research. For thick cloud removal methods, the concept of thick cloud removal methods is extended, various thick cloud removal methods are comprehensively summarized, the advantages and disadvantages of each method are analyzed, and the future research focus is prospected, which helps relevant researchers to have a comprehensive and clear understanding of this direction.
What problem does this paper attempt to address?