分层提取法在黄河三角洲湿地信息提取中的应用

On the Application of Hierarchical Extraction Method in the Yellow River Delta Wetland Information Extraction

黄河三角洲湿地是世界暖温带保存最广阔、最完善、最年轻的湿地生态系统,具有重要的研究与保护价值。本文以Landsat TM影像作为数据源,采用分层提取思想,利用人工目视解译与决策树分类,运用多种指数模型,提取了2004年黄河三角洲湿地信息。研究表明,采用分层提取法,能够较高精度地提取各种湿地信息,总体精度达到85.53%,其中浅海水域(100%)、养殖池塘(100%)、滩涂(87.5%)、建筑用地(84.62%)的分类精度较好,盐碱地(71.43%)和裸地(73.91%)的分类精度有待进一步的提高。

Wetland of Yellow River Delta is the vastest , the most perfect and the youngest wetland ecosystem in the warm temperate zone of world , so it has important research and conservation value .In this paper , we extract the Yellow River Delta wetland informa-tion of 2004 using hierarchical extraction method , taking advantage of manual visual interpretation and decision tree classification , and using a variety of index models .Research shows that hierarchical extraction method can extract various kinds of wetland information with high accuracy.The overall accuracy is 85.53%, with better accuracy of shallow waters （100%）, fish ponds （100%）, beach （87.5%）, and construction site （84.62%）, while the accuracy of saline -alkali soil （71.43%） and bare land （73.91%） needs to be further improved .