摘要
近年来,随着航空航天事业的高速发展,带动了遥感对地观测技术的进步,为高分影像的获取奠定了基础。作为地物类别中的主要内容和地形图中的重要成图元素,建筑物的识别与提取,直接影响到地物提取的自动化水平。因此,高分辨率遥感影像中建筑物的提取是图像处理领域中的主要研究内容之一。为了提高城市建筑物信息提取精度,本文改进了常规的面向对象方法,以航空遥感影像和SPOT-6影像为对象针对其下垫面结构复杂的特性,采用多尺度分割和多规则结合的方法自动提取建筑物信息,并通过样本区进行了精度验证,将提取的结果与传统分类方法所得到的结果相互比较。研究结果表明,面向对象的多尺度分割对高分影像中建筑物的提取具有较好地效果,KIA精度达到了0.76,为城市建筑物信息提取的应用提供了新思路。
In recent years the rapid development of aerospace enterprise drives the progress of remote sensing for earth observation technology,which laid a foundation for high image acquisition. As the main contents and important mapping elements in topographic map,the identification and extraction of buildings directly affects the automation level of feature extraction. Therefore the extraction of buildings in the high resolution remote sensing image is one of the main research contents in the field of image processing. This paper in order to improve the extraction accuracy of the city buildings and the conventional object- oriented method,using the combined method of multi- scale segmentation and many rules automatically extract information building based on aviation remote sensing image and SPOT- 6 images as objects for its characteristics of complex structure. Through the accuracy verification for the sample area,it is compared with the traditional classification method the results. The research results show that the object- oriented and multi- scale segmentation for extracting building from the high image has good effect,KIA accuracy reached 0. 76,and it provides a new way of thinking for the application of city building information extraction.
出处
《测绘与空间地理信息》
2016年第6期17-20,共4页
Geomatics & Spatial Information Technology
基金
黑龙江省自然科学基金(E201203)
黑龙江省教育厅科学技术研究项目(12541655)资助
关键词
高分影像
建筑物提取
图像分割
尺度参数
多尺度分割
high spatial resolution image
building extraction
image segmentation
scale parameter
multi-scale segmentation
