首页|New Robotics Findings from Southeast University Reported (Multiview Registratio n of Partially Overlapping Point Clouds for Robotic Manipulation)
New Robotics Findings from Southeast University Reported (Multiview Registratio n of Partially Overlapping Point Clouds for Robotic Manipulation)
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
Investigators publish new report on Ro botics. According to news reporting from Nanjing, People's Republic of China, by NewsRx journalists, research stated, "Point cloud registration is a fundamental task in intelligent robots, aiming to achieve globally consistent geometric str uctures and providing data support for robotic manipulation. Due to the limited view of measurement devices, it is necessary to collect point clouds from multip le views to construct a complete model." Funders for this research include National Natural Science Foundation of China ( NSFC), Ministry of Industry and Information Technology Basic Research Project. The news correspondents obtained a quote from the research from Southeast Univer sity, "Previous multi-view registration methods rely on sufficient overlap and r egistering all pairs of point clouds, resulting in slow convergence and high cum ulative errors. To solve these challenges, we present a multi-view registration method based on the point-to-plane model and pose graph. We introduce a robust k ernel into the objective function to diminish registration errors caused by mism atched points. Additionally, an enhanced Euclidean clustering method is proposed for extracting object point clouds. Subsequently, by establishing pose constrai nts on non-adjacent frames of point clouds, the cumulative error is reduced, ach ieving global optimization based on the pose graph. Experimental results demonst rate the robustness of our method with respect to overlap ratios, successfully r egistering point clouds with overlap ratio exceeding 30$% ."
NanjingPeople's Republic of ChinaAsiaEmerging TechnologiesMachine LearningRoboticsRobotsSoutheast Universi ty
NETL
NSTL
