基于CORONA KH-4B影像的湿地景观格局重建——以杭州西溪湿地为例

多源遥感影像是湿地制图和变化检测的重要依据，目前使用的商业高分辨率遥感影像仅能重现近20 a的湿地景观格局.利用解密的美国20世纪六七十年代CORONA KH-4B高分辨率间谍卫星影像，以杭州西溪湿地为例，利用面向对象分类方法进行湿地地物自动分类，进行湿地景观格局指数提取.结果表明，利用CORONA KH-4B影像和面向对象分类方法可有效进行湿地分类，重现近50 a前湿地景观格局.与近期Worldview 2影像和高分1号影像的提取结果的对比表明，50 a前西溪湿地主要呈耕地景观要素与池塘、河流等湿地景观要素共存的格局，后者在农业开发的影响下破碎度较高；现今西溪湿地区域优势景观要素为湿地与城市建筑，该演变的主要驱动力为杭州城西建设与西溪湿地综合保护工程的实施.
Abstract: Multi-source remote sensing images are important for wetland mapping and change detection. However, the commercial high-spatial-resolution satellite images currently used can only reproduce the wetland landscape pattern for nearly 20 years. The declassified CORONA KH-4B image acquired by American spy satellite in the 1960s and early 1970s was used for wetland mapping in Xixi Wetland in Hangzhou, China. The object-oriented classification method was used to classify wetland objects automatically, and wetland landscape pattern indexes were extracted. The result showed that the wetland classification and reconstruction of wetland landscape pattern nearly 50 years ago can be effectively carried out by CORONA KH-4B image and object-oriented classification method. Comparison of the landscape patterns extracted from recent Worldview 2 image and GF-1 image with that of CORONA KH-4B image showed that Xixi Wetland was mainly characterized by the coexistence of the cultivated land landscape element and the natural wetland landscape element 50 years ago, and the fragmentation degree of the latter was high under the influence of agricultural development. Today, the dominant landscape elements of Xixi Wetland are wetland and civic construction. The main driving forces of this evolution are the construction of the west part of the city and the implementation of the comprehensive protection project of Xixi Wetland.