基于高分一号宽幅多光谱影像的鄱阳湖湿地分类研究

发布时间：2018-12-20 08:01

【摘要】：典型湿地类型分类对于湿地生态环境保护和生态环境建设具有重要意义。本研究以高分一号宽幅多光谱影像数据作为数据源,以鄱阳湖典型湿地作为研究对象,进行典型湿地类型分类。研究采用主成分分析、归一化植被指数、比值植被指数和归一化差异水体指数等方法对不同湿地类型进行光谱特征分析,结果表明:高分一号宽幅多光谱影像反演的归一化植被指数和归一化水体指数能够较好地区分水体、泥滩、挺水植物和湿生植物等常见湿地类型。通过归一化植被指数将湿地类型分为植被和非植被,归一化差异水体指数将植被类型进一步分为挺水植物和湿生植物,将非植被分为水体和泥滩。构建决策树对典型湿地类型进行自动分类,经过精度评价和分析验证,该方法针对高分一号宽幅数据进行典型湿地分类总体精度能够达到82.28%,Kappa系数0.7346,优于常规的监督分类和非监督分类。
[Abstract]:The classification of typical wetland types plays an important role in wetland ecological environment protection and ecological environment construction. In this study, the data of wide-amplitude multispectral image of Gaofeng No. 1 were used as data source, and the typical wetland of Poyang Lake was used as the research object to classify the typical wetland types. The spectral characteristics of different wetland types were analyzed by principal component analysis, normalized vegetation index, ratio vegetation index and normalized differential water body index. The results show that the normalized vegetation index and normalized water body index can be used to distinguish common wetland types such as water body, mudflat, occultation plant and hygrophyte. The wetland types are divided into vegetation and non-vegetation by normalized vegetation index, and the vegetation types are further divided into octopus plants and wet plants by normalized difference water body index, and non-vegetation is divided into water body and mudflat. The decision tree is constructed for automatic classification of typical wetland types. Through precision evaluation and analysis, the overall accuracy of typical wetland classification can reach 82.28% Kappa coefficient 0.7346. It is superior to the conventional supervised classification and unsupervised classification.
【作者单位】：黑龙江第二测绘工程院;江西省环境保护科学研究院;
【分类号】：P237

【相似文献】