摘要
基于曲率变化的曲线分段去噪法,先读入数据,将点云数据按扫描线顺序存储到数组A。再对除首末点外位于同一扫描线上的点,计算各点曲率。接着计算数组A中,等间距两点的曲率差,放到数组B,提取数组B中大于给定阈值的n个点。然后拟合曲线,计算各点到曲线的距离,并删除距离超差的点,直到处理完n+1段点云,由此处理完所有扫描线。
The curve subsection noise remove algorithm based on curvature variation, read data at first, and store cloud-data in array A according to scanning line sequence. Then calculate curvature of each point on the same scanning line except start and end points. After that, calculate the curvature of equal space between two points and send them to array B and extract N points which are more than given threshold value. Then, simulate curve, calculate the distance between each point and curve. Delete the error distance point. At last, deal with N＋I subsection cloud data and scannin~ line.
出处
《兵工自动化》
2007年第10期39-40,共2页
Ordnance Industry Automation
关键词
逆向工程
噪声点
曲率变化
Reverse engineering
Noise point
Curvature varying
