基于形态学建筑物指数的城市建筑物提取及其高度估算

发布时间：2017-12-31 06:42

本文关键词：基于形态学建筑物指数的城市建筑物提取及其高度估算　出处：《遥感技术与应用》2015年01期 　论文类型：期刊论文

【摘要】：通过引入高分辨率影像的形态学建筑物指数和阴影指数,并结合面向对象的地物信息提取思想,准确地提取出城市建筑物及其阴影,进而实现了城市建筑物的高度估算。首先,利用形态学建筑物指数的多方向多尺度特征,将建筑物与邻近光谱相似的道路目标进行分离;其次,采用双阈值策略提取建筑物与相应的阴影,进一步提高了建筑物的提取精度;最后,根据成像时刻卫星和太阳的高度角、方位角,建立建筑物阴影长度与建筑物高度的估算模型。试验利用厦门市思明区软件园资源三号(ZY-3)数据进行城市建筑物提取及其高度估算,证实该方法能够较准确地估算出建筑物的高度信息,并且比基于SVM的监督分类方法具有更高的建筑物提取精度,建筑物高度估算的中误差可达±1m。
[Abstract]:By introducing the morphological building index and shadow index of high-resolution image and combining the idea of object oriented feature information extraction, the urban buildings and their shadows are extracted accurately. Then the height estimation of urban buildings is realized. Firstly, the multi-direction and multi-scale characteristics of morphological building index are used to separate the buildings from the road targets with similar spectra. Secondly, the double threshold strategy is used to extract the building and the corresponding shadow, which further improves the precision of the building extraction. Finally, the azimuth and azimuth of the satellite and the sun according to the imaging time. The estimation model of shadow length and height of buildings was established. The data of software park No. 3 of Siming District of Xiamen City were used to extract and estimate the height of urban buildings. It is proved that this method can accurately estimate the height information of buildings, and has a higher accuracy than the supervised classification method based on SVM, and the median error of building height estimation can reach 卤1 m.
【作者单位】：福州大学空间数据挖掘与信息共享教育部重点实验室;军事医学科学院;
【基金】：国家863计划项目(2012AA022007)
【分类号】：TU198
【正文快照】： 2.军事医学科学院,北京100071)1引言随着遥感技术的不断进步,光学卫星影像的空间分辨率得到极大提高,高分辨率影像上地物的空间几何、结构、形状、纹理和细节更明显,使得研究城市内部建筑物分布及其高度成为了可能。查勇等[1]提出一种归一化建筑物指数NDBI,利用TM影像的第5、

【相似文献】