宁夏哈巴湖国家级自然保护区植物群落遥感分类研究

Remote Sensing Classification of Plant Communitity in Haba Lake National Nature Reserve of Ningxia

为准确监测宁夏哈巴湖国家级自然保护区植物资源,以2017年Landsat8遥感影像、高分卫星遥感影像结合实地调查和无人机多光谱影像,分析保护区9种典型植物群落的光谱特征,并利用2种分类方法对保护区植被进行分类。研究结果表明,面向对象分类法分类精度为79.90%,最大似然法分类精度仅为52%,面向对象分类方法比最大似然法更适合进行研究区植物群落尺度的分类制图;研究区分布范围广泛且集中连片分布的是沙柳群落和柠条群落,与多年来保护区长期治理措施有关。

In order to accurately monitor the plant resources of Haba Lake National Nature Reserve in Ningxia and provide scientific basis for the ecological protection of the reserve,the spectral characteristics of 9 typical plant communities in the reserve were analyzed based on Landsat 8 remote sensing images and high-resolution satellite remote sensing images in 2017,combined with field investigation and UAV multispectral images,Two classification methods were used to classify the vegetation in the reserve.The results showed that the classification accuracy of object-oriented classification method was 79.90%,and the classification accuracy of maximum likelihood method was less than 52%;Salix psammophila community and Caragana korshinskii community were concentrated and widely distributed in the study area,which was related to the long-term management measures of the reserve over the years.