【摘要】 为精准识别水体信息并实时监测湖泊水体时空特征及其环境特征变化情况，以洞庭湖为例，基于Landsat-8影像数据分别使用改进的归一化差异水体指数（Modified Normalized Difference Water Index, MNDWI）、自动水体提取指数(Automated Water Extraction Index, AWEI_sh)、支持向量机（Support Vector Machine, SVM）、人工神经网络（Artificial Neural Networks, ANNs）、随机森林（Random Forest, RF）等5种方法提取枯、丰水期水体分布信息，通过精度指标评价及影响因素分析，旨在找到提取精度高、鲁棒性强的水体提取方法。结果表明：5种方法中SVM法水体提取总精度最高且泛化能力良好。研究成果可为各方法适用性提供一定参考，并通过定量分析揭示漏提率在提取精度评价指标中的重要性。更多还原

【Abstract】 In order to accurately identify the water information and monitor the spatial and temporal characteristics of lake water and the changes of their environmental characteristics in real time, taking Dongting Lake as an example, five methods including Modified Normalized Difference Water Index（MNDWI）,Automated Water Extraction Index(AWEI_sh),Support Vector Machine（SVM）,Artificial Neural Networks（ANNs） and Random Forest（RF）,were used to extract water distribution information during the dry and wet seasons based on Landsat-8 image data.By comparing the extraction accuracy and analyzing the influencing factors, the aim is to find a water extraction method with high extraction accuracy and robustness.The results show that the SVM method has the highest overall accuracy and good generalization ability among the five methods.This study can provide a reference for the applicability of each method and reveal the importance of the omission rate in the evaluation index of extraction accuracy through quantitative analysis.更多还原