An improved surface water extraction method by integrating multi-type priori information from remote sensing

• DOI: 10.1016/j.jag.2023.103529
• Publication Date: 2023-10-23T20:42:43Z
• Authors: Bingyu Zhao; Jianjun Wu; Xinyi Han; Feng Tian; Mengxue Liu; Meng Chen; Jingyu Lin
• Keywords: Robustness; Neighbourhood (mathematics)

Surface water mapping based on historical, neighbourhood, and other priori information has shown improved accuracy. However, the accuracy can be compromised due to lack of consideration for dynamics in proximity period limited utilization quantitative methods integrating multiple types information. In this study, an unsupervised surface extraction method that integrates proximity, neighbourhood from remote sensing is proposed enhance The experiments were conducted Poyang Lake, a region characterized by active hydrological phenomena. Coarsely extracted extents using OTSU all available Sentinel-1/2 images within one-month current moment utilized as estimate probability (WP). optimized WP was then obtained combining estimated probabilities Bayesian Model Averaging (BMA) method. experimental results demonstrate outperforms traditional body such K-Means, IsoData, terms Moreover, integration BMA higher compared each separately. Specifically, incorporation significantly enhances estimation WP. summary, offers high accuracy, automation, strong robustness, making it applicable areas complex changes.