大型复杂锻铸件三维扫描点云数据去噪及光顺研究

Study on denoising and smoothing method of 3D-scaning point cloud data to the large complex malleable cast parts

逆向工程,对锻铸件的三维点云数据处理是获得复杂零件外形尺寸的关键环节。如何从散乱的、多噪点的点云数据中得到无噪点、光滑的三维点云是人们一直研究的重点。依据散乱点云中的噪声点的特征进行分类,并提出一种分步去噪及光顺方法。首先用格网阀值法去除第一类噪点,用格网法去除第二类噪点,最后用双边滤波法对数据进行光顺处理。结果表明该方法不仅有效去除了锻铸件的三维点云数据的噪声点,同时获得了光滑的测量表面,保留了尖锐特征。

Reverse engineering is the key part to obtain dimensions of complex parts for the processing of three- dimensional point cloud data of malleable cast parts. The key point is how to get the noiseless and smooth 3D point cloud from the scattered and multi-noise point cloud data. According to characteristics of noise points in the scattered cloud, this paper presents two methods： step-by-step denoising and smoothing method. First, the first kind of noise is removed by using grid threshold method. Second, the second kind of noise is removed by using virtual grid method. Finally, the above data is smoothed by using bilateral filtering method. The results show that the method not only removes the noisy points of 3D point cloud data, but also obtains a smooth meas- urement of surface and retains the characteristics of sharp features.