摘要
为了从高分辨率遥感图像中完整提取建筑物区域,采用区域分割的原理,研究了建筑物自动检测的方法。该方法首先利用利用K-Mean分类方法将地物分为两类:人工地物类和非人工地物类,然后利用阴影、Mean Shift分割信息来剔除人工地物类中干扰区域,再根据形状分析来确定真实的建筑物区域。本文用上述方法对高分辨率航空影像进行了实验,实验结果证明了该方法有着较高的识别率、较好的准确性和鲁棒性,具有实用价值。
In order to extract building targets in high resolution remote sensing images automatically,a new method for automatic building detection in color aerial images based on region segmentation is proposed in this paper.Firstly,an unsupervised classification based on simple K-Mean clustering is used to classify the color aerial images.The image is classified into two classes of 1） man-made objects and 2） non-man-made objects.Secondly,the shadow information and Mean Shift segmentation are used to separate the false regions from man-made objects.Finally,the true regions with buildings are obtained by shape analysis.A case study was conducted to apply the method proposed in this paper on high resolution aerial images.The experimental results demonstrate that this method has a high precision and rational robustness.
出处
《辽宁工程技术大学学报（自然科学版）》
CAS
北大核心
2010年第6期1058-1061,共4页
Journal of Liaoning Technical University (Natural Science)
基金
国家自然科学基金资助项目(40901195)
中国博士后科学基金项目(20070420412)
国家863基金资助项目(2007AA12Z215
2007AA12Z333)
地理空间信息工程国家测绘局重点实验室基金项目(200830)
公益性科研院所基本科研业务费专项资金项目(77734
77722)
关键词
区域分割
建筑物自动检测
信息融合
region segmentation
automatic building detection
information fusion
