基于双边滤波算法的树木点云模型去噪方法研究  

Method on tree point cloud model de-noising based on bilateral filtering

作　　者：管西鹏 陈宇拓[1] 张怀清[2] 刘闽[2] 李永亮[2] 

机构地区：[1]中南林业科技大学计算机与信息工程学院,湖南长沙410004 [2]中国林业科学研究院资源信息所,北京100091

出　　处：《中南林业科技大学学报》2015年第9期83-87,共5页Journal of Central South University of Forestry & Technology

基　　金：国家863计划资助项目(2012AA102002)

摘　　要：针对三维激光扫描仪获取的树木点云数据存在噪声点的问题,结合双边滤波算法,通过设定空间邻近度因子σc值不变、灰度邻近度因子σs变化以及σs值不变、σc变化的2种不同参数取值状态,根据计算得到的双边滤波权因子α,选择剔除部分与当前采样点距离过大的相邻采样点,更新所有的数据点,最后得到经过滤波后的新数据点。经过对比分析2种参数取值状态下的不同去噪结果,可得到合理的参数取值,进而实现对树木点云模型去噪的处理,且达到保持原特征的目的。最后选取2个合适的实验数据模型,进一步对实验结果进行验证。实验结果表明该算法简单高效,噪声去除效果良好,具有较好的实用性和研究意义。For the problems of noises in trees point cloud data obtained by a three-dimensional laser scanner, by combining bilateral ifltering algorithm, through two kinds of states in parameter values were set thatthe spatial proximity factor valueσc unchanged, while gray proximity factorσs change andσs value unchanged, whileσs change. According to the bilateralifltering weight factors, the adjacent sampling points were removed because it is too large with the current sampling point and update all data points. Afterifltering, the new data points were gotten. After comparatively analyzing the de-noisingresults of the two different parameter values state, the reasonable parameter values were obtained and thus realizing the purpose of the de-noising processing in trees point cloud model and maintaining the original features. Finally, two suitable experimental data models were selected, and the experimental results were validated further. The experimental results show that the algorithm is simple and efifcient and noise removal works well. It has good practicability and signiifcance.

关 键 词：树木测量 双边滤波算法 点云去噪 空间邻近 灰度邻近 

分 类 号：S758[农业科学—森林经理学]