基于直线段拟合的遥感影像居民区轮廓规则化方法  

A method of residential area contours regularization in remote sensing image based on straight line segment fitting

作　　者：赵明瑜[1,2] 刘松林 徐国庆[3] 杨苗苗 ZHAO Mingyu;LIU Songlin;XU Guoqing;YANG Miaomiao(Xian Research Institute of Surveying and Mapping,Xi'an 710054,China;State Key Laboratory of Geo-Information Engineering,Xi'an 710054,China;Technical Division of Surveying and Uapping,Xi'an 710054,China)

机构地区：[1]西安测绘研究所,陕西西安710054 [2]地理信息工程国家重点实验室,陕西西安710054 [3]西安测绘信息技术总站,陕西西安710054

出　　处：《测绘科学与工程》2019年第3期29-33,共5页Geomatics Science and Engineering

摘　　要：针对遥感影像智能提取结果中居民区轮廓提取结果不规则,具有冗余性等问题,提出了一种新的居民区轮廓规则化方法。首先对提取的栅格数据二值化,利用数学形态学算法提取二值化的区域边界,获取居民区边界轮廓线;然后采用折线法思想对边界轮廓进行直线段拟合;最后对拟合后结果进行智能综合优化,得到规则化后的居民区轮廓。试验证明,该方法可以有效地对居民区轮廓进行规则化处理,尤其适用于由大量栅格点表示,轮廓线十分曲折的轮廓,且方法简单,易于算法设计与程序实现,对遥感影像居民区智能提取结果优化、精化效果明显。In order to solve the problems of irregular extraction and redundancy of residential area contours in the intelligent extraction of remote sensing images,a new method of residential area contour regularization is proposed.Firstly,the extracted ras-ter data is binarized,and the boundary of the binarized region is extracted using mathematical morphology algorithm to obtain the boundary contours of residential areas.Then the polyline method is used to fit the boundary contours to the straight line segment.Finally,the results after fitting are intelligently integrated to obtain the residential area contours after regularization.The test proves that the method can effectively regularize the outline of the residential area,especially winding contours with a large num-ber of grid points.The method is simple and convenient for algorithm design and program implementation,and the optimization and refinement of the intelligent extraction of residential areas in remote sensing images are effective.

关 键 词：直线段拟合 居民区 轮廓 规则化 形态学算法 

分 类 号：P237[天文地球—摄影测量与遥感]