{"title":"GFFS: Gravitational-Force Feature Real Time Segmentation of 3D Point Cloud","authors":"Chunhao Shi, Chunyang Wang, Xuelian Liu, Boyang Xiao, Shaoyu Sun, Guan Xi, Wenqian Qiu","doi":"10.1109/IPEC54454.2022.9777336","DOIUrl":null,"url":null,"abstract":"Segmentation of point cloud is a momentous technology for autonomous vehicles. Due to the diversity and complexity of road environment, instance segmentation based on point cloud is a challenging task. Addressing this challenge, a gravitational-force feature segmentation of 3D point cloud (GFFS) is proposed. Firstly it divides the point into areas, secondly it sets the radius of the spherical neighborhood corresponding to each area based on the size and density of point cloud, thirdly it calculates the gravitational forces between the space of each spherical neighborhood and the barycenter of the area. GFFS performs instance segmentation of all kinds of objects in point cloud scene by analyzing the magnitude and continuity of gravitational forces. Finally, the superiority and effectiveness of the algorithm are verified by experiments.","PeriodicalId":232563,"journal":{"name":"2022 IEEE Asia-Pacific Conference on Image Processing, Electronics and Computers (IPEC)","volume":"48 3","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2022-04-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2022 IEEE Asia-Pacific Conference on Image Processing, Electronics and Computers (IPEC)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/IPEC54454.2022.9777336","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
Segmentation of point cloud is a momentous technology for autonomous vehicles. Due to the diversity and complexity of road environment, instance segmentation based on point cloud is a challenging task. Addressing this challenge, a gravitational-force feature segmentation of 3D point cloud (GFFS) is proposed. Firstly it divides the point into areas, secondly it sets the radius of the spherical neighborhood corresponding to each area based on the size and density of point cloud, thirdly it calculates the gravitational forces between the space of each spherical neighborhood and the barycenter of the area. GFFS performs instance segmentation of all kinds of objects in point cloud scene by analyzing the magnitude and continuity of gravitational forces. Finally, the superiority and effectiveness of the algorithm are verified by experiments.