摘要
随着成像技术的不断进步,人们对空间三维物体的识别、分割的研究逐渐深入。在计算机中储存三维物体方式中,点云凭借自身优良的特性,广泛应用于相关研究。外界因素影响如传感器精度、环境光线、物体表面反射等,会导致原始获取的点云数据存在噪点、冗余、稀疏等问题,因此研究点云相关算法前需要预处理点云,从而减少无关因素对算法性能的影响。
With the continuous progress of imaging technology,people have gradually deepened their research on the recognition and segmentation of three-dimensional spatial objects.In the way of storing three-dimensional objects in computers,point clouds are widely used in related research due to their excellent characteristics.External factors,such as sensor accuracy,ambient light,and object surface reflection,can lead to issues such as noise,redundancy,and sparsity in the originally obtained point cloud data.Therefore,point clouds need to be preprocessed before studying point cloud related algorithms to reduce the impact of unrelated factors on algorithm performance.
作者
樊轶铖
张俊琪
崔宸
刘彦松
简春
FAN Yicheng;ZHANG Junqi;CUI Chen;LIU Yansong;JIAN Chun(Nuclear Power Institute of China,Chengdu Sichuan 610000,China)
出处
《信息与电脑》
2023年第6期206-209,共4页
Information & Computer
关键词
点云
预处理
采样
特征提取
point cloud
preprocessing
sampling
feature extraction
