文献类型：期刊文献

英文题名：Research on Shallow Water Depth Remote Sensing Based on the Improvement of the Newton-Raphson Optimizer

作者：Li, Yanran[1];Liu, Bei[2];Chai, Xia[2];Guo, Fengcheng[1];Li, Yongze[2];Fu, Dongyang[2,3,4]

机构：[1]Guangdong Ocean Univ, Coll Oceanog & Meteorol, Zhanjiang 524088, Peoples R China;[2]Guangdong Ocean Univ, Coll Elect & Informat Engn, Zhanjiang 524088, Peoples R China;[3]South China Sea Resources Big Data Ctr, Southern Marine Sci & Engn Guangdong Lab Zhanjiang, Zhanjiang 524025, Peoples R China;[4]Guangdong Engn Technol Res Ctr Marine Remote Sensi, Zhanjiang 524088, Peoples R China

年份：2025

卷号：17

期号：4

外文期刊名：WATER

收录：SCI-EXPANDED(收录号：WOS:001429502500001)、、EI(收录号：20250917974993)、Scopus(收录号：2-s2.0-85219062181)、WOS

基金：This research was funded in part by the National Key Research and Development Program of China under grant no. 2022YFC3103101, Key Special Project for Introduced Talents Team of Southern Marine Science and Engineering Guangdong Laboratory (GML2021GD0809), National Natural Science Foundation of China (No. 42206187), Key projects of the Guangdong Education Department (2023ZDZX4009).

语种：英文

外文关键词：water depth inversion; Newton-Raphson optimizer; XGBoost model; Sentinel-2B

外文摘要：The precise acquisition of water depth data in nearshore shallow waters bears considerable strategic significance for marine environmental monitoring, resource stewardship, navigational infrastructure development, and military security. Conventional bathymetric survey methodologies are constrained by their spatial and temporal limitations, thus failing to satisfy the requirements of large-scale, real-time surveillance. While satellite remote sensing technologies present a novel approach to water depth inversion in shallow waters, attaining high-precision inversion in nearshore areas characterized by elevated levels of suspended sediments and diminished transparency remains a formidable challenge. To tackle this issue, this study introduces an enhanced XGBoost model grounded in the Newton-Raphson optimizer (NRBO-XGBoost) and successfully applies it to water depth inversion investigations in the nearshore shallow waters of the Beibu Gulf. The research amalgamates Sentinel-2B multispectral imagery, nautical chart data, and in situ water depth measurements. By ingeniously integrating the Newton-Raphson optimizer with the XGBoost framework, the study realizes the automatic configuration of model training parameters, markedly elevating inversion accuracy. The findings reveal that the NRBO-XGBoost model attains a coefficient of determination (R2) of 0.85 when compared to nautical chart water depth data, alongside a scatter index (SI) of 21%, substantially surpassing conventional models. Additional validation analyses indicate that the model achieves a coefficient of determination (R2) of 0.86 with field-measured data, a mean absolute error (MAE) of 1.60 m, a root mean square error (RMSE) of 2.13 m, and a scatter index (SI) of 13%. Moreover, the model exhibits exceptional performance in extended applications within the waters of Zhanjiang Port (R2 = 0.90), unequivocally affirming its dependability and practicality in intricate nearshore water environments. This study not only provides a fresh solution for remotely sensing water depth in complex nearshore water settings but also imparts valuable technical insights into the associated underwater surveys and marine resource exploitation.