基于遥感图像的建筑物提取.pdfVIP

摘要 为了创建”数字城市 “系统，同时也为了的目标识别和提取，遥感图像或 空影 像中建筑物提取作为其主要组成部分已经备受关注。本文在当今复杂场景图像中的建 筑物提取技术上，重点研究简单城区建筑物的提取技术,根据两种不同类型的城区建筑 物影像,依次给出了较合理的图像预处理算法。 一般复杂城区影像中的建筑物区域提取是为了缩小建筑物特征点的检测、识别和 匹配的搜索范围,同时改善建筑物的目标识别以及增效三维重建；同时也起到了创建建 筑物的纹理库的作用,为进一步完成建筑物的纹理的自动贴图功能做出不懈努力。本文 另辟蹊径，重点设计研究彩色特征建筑群的分类提取，进一步在准确提取图像中的建 筑物的基础上更有效排除背景干扰信息。 本文主要考虑使用RGB 向量空间分割进行建筑物的识别和提取,在充分考虑和利用 建筑物屋顶的彩色特征化的同时,根据色调聚类分析，以及数学形态学的操作处理，对 具有明显特征化的建筑物进行提取，使其在几何特征和位置特征方面都更接近实际建 筑模型。 关键字 ：建筑物提取，色调聚类，RGB 向量空间分割，遥感图像 Abstract In order to create the digital city system, but also for the target recognition and extraction, remote sensing images or aerial images of building extraction as its main component has received much attention.This article in todays complex scene image building extraction technology, focuses on simple urban building extraction technology, according to two different types of buildings in urban area of image, which in turn gives the reasonable image pre processing algorithms. General complex urban image building area extraction is to narrow the buildings feature point detection, identification and matching the search range and at the same time, improve the building of target recognition and efficiency of 3 reconstruction; also played a creating texture library buildings, for a complete building texture automatic mapping function to make unremitting efforts.This another way, focusing on the design of color characteristics of buildings classification and extraction of further buildings in accurately extracting image based on more effectively eliminate interference of background information. In this paper, we mainly consider using RGB vector space partition for the recognition and extraction of the building, take account of and make use of the color feature of the roof of a building of at the same time, according to the hue clustering analysis,