北京测绘2025，Vol.39Issue(6)：873-881,9.DOI:10.19580/j.cnki.1007-3000.2025.06.020
机器学习协同多源遥感数据融合的红树林冠层高度估测
Mangrove canopy height estimation based on machine learning collaboration with multi-source remote sensing data fusion
摘要
Abstract
In response to the issues of poor accessibility,low efficiency,and difficulty in conducting large-scale monitoring in traditional mangrove plot surveys due to the complex terrain of the intertidal zone,this paper integrated unmanned aerial vehicle(UAV)and multi-source remote sensing data and adopted three machine learning methods,namely extreme gradient boosting(XGBoost),gradient boosting decision tree(GBDT),and random forest(RF),so as to construct a prediction model for the canopy height of mangroves in Hainan Province.The results show that the GBDT algorithm performs the best,with a modeling coefficient of determination R2 of 0.89,a model bias B of 0.01 m,a relative bias Br of 0.42%,a root mean square error(RMSE)E of 1.39 m,and a relative root mean square error(RRMSE)Er of 39.84%.The prediction results show that the average canopy height of mangroves in Hainan Province increases from 4.24 m in 2019 to 4.82 m in 2022,with an overall trend of continuous growth.By combining multi-source remote sensing data and machine learning algorithms,an efficient mangrove canopy height monitoring model is constructed,which provides strong data support and technical reference for the management and protection of mangroves in Hainan Province and further proves the great potential of remote sensing technology in ecological monitoring of mangroves.关键词
红树林/冠层高度/遥感监测/多源数据融合/机器学习Key words
mangrove/canopy height/remote sensing monitoring/multi-source data fusion/machine learning分类
天文与地球科学引用本文复制引用
尹招德,冯仲科..机器学习协同多源遥感数据融合的红树林冠层高度估测[J].北京测绘,2025,39(6):873-881,9.基金项目
国家自然科学基金(32160364) （32160364）
