基于滚动圆的点云数据边缘检测算法  

Edge detection algorithm for point cloud based on rolling-circle

作　　者：李宝顺[1] 岑红燕[1] 包亚萍[1] 李义丰[1] 

机构地区：[1]南京工业大学电子与信息工程学院,江苏南京211816

出　　处：《计算机工程与设计》2013年第5期1836-1840,共5页Computer Engineering and Design

基　　金：国家自然科学基金项目(11104142;61001077)

摘　　要：针对点云数据边缘较难检测这一问题,提出了一种快速有效的点云数据边缘检测新算法。该算法以点云数据的全局特性为基础,通过分析三维空间内平面几何曲线的全局边缘性质,构建了基于滚动圆的边缘检测新模型。采用滚动圆的方法针对不同边缘类型进行相应建模,定义了多个特征系数并基于此建立边缘判定准则,实现快速边缘定位。利用公共数据库进行多组实验,分析结果表明,该方法能快速稳定地检测物体边缘,具有较高的检测质量和速度性能。Aiming at the problem that edge of point cloud is difficult to detect, a novel and effective edge detection algorithm for point cloud data is presented. By analyzing global properties of point cloud data and explaining global edge properties of planar geometric curve in three dimensions, a new edge detection model based on rolling-circle is developed. Different edge types are characterized by different models. In order to realize rapid edge localization, several characteristic coefficients are defined for edge criterion. Many experiments with public database show this algorithm can detect objects＇ edges fast and stably, also can achieve good detection quality and high-speed performance.

关 键 词：点云数据 滚动圆 边缘检测 全局特性 边缘判定准则 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]