摘要
针对二维激光扫描仪获取的原始点云数据存在噪声点而影响测量精度的问题,提出了基于先验概率算法与曲线概率阈值分割的点云数据消噪方法。以异形截面环形件为实验对象,运用该消噪方法对原始数据进行处理,分别采用原始数据和消噪数据计算了环件台阶轴的高度,并将其与环件的真实值进行了对比。对比结果表明,消噪后数据的计算值更接近环件的真实值,从而验证该算法能有效提高测量精度。
The original point cloud data acquired from two dimensional laser scanner had noises, and the noises affected measurement accuracy. For this problem, a noise cancellation algorithm method combinated priori probability with curve probability threshold segmentation was proposed. Two di-mensional laser scanner scaned the profiled annular section, and was used as experimental object, then acquired original point cloud data .The proposed denoising method dealt with the original point cloud data and got noise cancellation data. The original point cloud data and noise cancellation data were used to calculate the heights of the ring steps. Comparisons of calculated values and real values show that the calculated values for data denoising are closer to the true values of the ring. The experimental results prove that the noise cancellation method may effectively improve measurement accuracy.
出处
《中国机械工程》
EI
CAS
CSCD
北大核心
2017年第8期936-940,945,共6页
China Mechanical Engineering
基金
国家自然科学基金资助项目(51675469)
河北环境工程学院科研基金资助项目(BJ201603)
关键词
二维激光扫描仪
点云消噪
先验概率
曲线概率阈值分割
异形截面环形件
外部轮廓线
two dimensional laser scanner
point cloud denoising
priori probability
probability threshold curve segmentation
profiled annular section
outer contour line
