面向对象的形态学建筑物指数及其高分辨率遥感影像建筑物提取应用

摘要

Building extraction from high resolution remote sensing images is a hot research topic in the field of photogrammetry and remote sensing.In this article, an object-based morphological building index (OBMBI) is constructed based on both image segmentation and graph-based top-hat reconstruction, and OBMBI is used for building extraction from high resolution remote sensing images.First, bidirectional mapping relationship between pixels, objects and graph-nodes are constructed.Second, the OBMBI image is built based on both graph-based top-hat reconstruction and the above mapping relationship.Third, a binary thresholding is performed on the OBMBI image, and the binary image is converted into vector format to derive the building polygons.Finally, the post-processing is made to optimize the extracted building polygons.Two images, including an aerial image and a panchromatic satellite image, are used to test both the proposed method and classic PanTex method.The experimental results suggest that our proposed method has a higher accuracy in building extraction than the classic PanTex method.On average, the correctness, the completeness and the quality of our method are respectively 9.49%, 11.26% and 14.11% better than those of the PanTex.%高分辨率遥感影像建筑物提取是摄影测量与遥感领域的一个热门研究主题.本文综合利用影像分割、基于图的数学形态学top-hat重建技术,提出了面向对象的形态学建筑物指数OBMBI,并将其应用于高分辨率遥感影像建筑物提取.首先,建立像素-对象-图节点的双向映射关系;然后,基于图的白top-hat重建和上述映射关系来构建OBMBI图像;接着,对该OBMBI图像二值化、矢量化以获取建筑物多边形;最后,对结果进行后处理优化.使用一景航空、一景卫星全色影像对本文方法和PanTex方法进行性能测试.试验表明,本文方法的建筑物提取精度显著的优于PanTex方法.其中,本文方法平均比PanTex方法的正确率高9.49%、完整率高11.26%、质量高14.11%.