基于植被指数模型的淡水湖泊湿地景观信息提取

高邮湖湿地是江苏省重要湿地之一,对生态、环境控制、调节气候和保护生物多样性具有重要意义。采用2007年的LandsatTM影像作为遥感信息源,选择影像的光谱特征和比值植被指数(RVI)、差值植被指数(DVI)、归一化植被指数(NDVI)、归一化差异绿度指数(NDGI)、土壤调节植被指数(SAVI)和最佳土壤调节植被指数(OSAVI)6种植被指数做了光谱特征分析,从而确定出最佳指数模型,并基于决策树方法,实现研究区景观信息的遥感分类。研究结果表明,决策树分类法易于综合多种特征进行遥感影像分类,植被指数参与到决策树分类中能够提高分类的总体精度,其总体精度达到79.58%,Kappa系数为0.721 0,分类结果理想且人工参与灵活。

Extraction of Landscape Information in Freshwater Lake Wetlands Based on Vegetation Index Model

Gaoyou Lake wetland is one of important wetlands in Jiangsu Province. It is of great importance in ecological, environmental control, climate regulation and biodiversity conservation. In this paper, Landsat Thematic Mapper (TM) data acquired in 2007 were used as data sources. Spectral characteristics and six vegetation indices including Ratio Vegetation Index(RVI), Differential Vegetation Index(DVI), Normalized Difference Vegetation Index(NDVI), Normalized Difference Green Index(NDGI), Soil Adjusted Vegetation Index(SAVI), Optimized SAVI(OSAVI), were used for analysis of spectral characteristics to determine the optimal vegetation index model. The landscape pattern classification was achieved with decision tree. The results show that multiple features are easy to be integrated by the decision tree for remote sensing images classification. Vegetation indices used in decision tree can increase the overall accuracy of classification. The overall accuracy of classification is 79.58% with high Kappa Coefficient of 0.721 0. The classification results are satisfying. It is flexible for human participation.