基于形态学的散乱点云轮廓特征线提取  

Morphology-based Scattered Point Cloud Contour Extraction

作　　者：程效军[1] 方芳[1] 

机构地区：[1]同济大学测绘与地理信息学院,上海200092

出　　处：《同济大学学报（自然科学版）》2014年第11期1738-1743,共6页Journal of Tongji University:Natural Science

基　　金：国家自然科学基金(40971241)

摘　　要：采用逆向工程领域切片技术和数学形态学方法提取点云的轮廓特征线.空间散乱点云密度高且无拓扑关系,通过切片分层可将空间点云转换为不同层的平面切片点云.借鉴图像处理方法,将切片点云转换成二值图像,使用形态学运算提取其轮廓像素,将轮廓像素转换为轮廓特征点并采用B样条曲线拟合成轮廓特征线.实验验证该方法可以得到高质量轮廓线,并可以有效地解决"多环"切片难以正确提取轮廓特征的问题.Slicing technology and morphological operators are employed to extract contour feature from scattered point cloud. Through slicing, unorganized point with high density can be transformed into point sets on a series of plane layers. Raster structure of digital image is utilized to organize sliced point cloud to generate binary images. Then morphological operators are employed to extract contour pixels. The contour pixels are transformed back to contour points, and B spline fitting is done to generate contour feature. The experiment results show that this method can produce high quality result and overcome the multi-loop problem.

关 键 词：切片技术 形态学 轮廓特征 散乱点云 

分 类 号：P232[天文地球—摄影测量与遥感]