Title :
High precision pose measurement for humanoid robot based on PnP and OI algorithms
Author :
Chen, Guodong ; Xu, De ; Yang, Ping
Author_Institution :
Inst. of Autom., Chinese Acad. of Sci., Beijing, China
Abstract :
In this paper, a method combining PnP and OI algorithms is developed to measure the pose of a humanoid robot in high precision in order to play table tennis. The PnP-based algorithm is employed to obtain the rough pose, which is taken as the initial value of the OI algorithm. The OI algorithm optimizes the result in order to ensure the orthogonal orientation matrix of the pose. Considering the real-time, the accuracy, and the robustness of the pose measurement, a sort of fast feature extraction algorithm including edge detection, line extraction and intersection computation is adopted to get the four corners in high precision. The experiments and results verify the effectiveness of the proposed method.
Keywords :
edge detection; feature extraction; humanoid robots; matrix algebra; pose estimation; robot vision; sport; OI algorithm; PnP algorithm; edge detection; feature extraction algorithm; high precision pose measurement; humanoid robot; intersection computation; line extraction; orthogonal iteration algorithm; orthogonal orientation matrix; perspective-n-point method; table tennis; Cameras; Detectors; Feature extraction; Humanoid robots; Image edge detection; Pixel; Robot vision systems; OI; PnP; Pose measurement; RANSAC; feature extraction; humanoid robot;
Conference_Titel :
Robotics and Biomimetics (ROBIO), 2010 IEEE International Conference on
Conference_Location :
Tianjin
Print_ISBN :
978-1-4244-9319-7
DOI :
10.1109/ROBIO.2010.5723397