Abstract
Based on the characteristics of large amount of computation, slow running speed and poor real-time quality in virtual reality pose estimate, a monocular camera-based image recognition method is proposed to obtain the spatial pose data of the object to be tracked efficiently. Mean-Shift algorithm is used to solve the problem of dynamic blur caused by high-speed rotation of objects in images. By obtaining the a priori shape data and size of the object to be tracked, so on as the constraint condition, the coordinates of the points marked by ArUco and the camera data of the calibrated data are obtained, and the three-dimensional space pose of the object to be tracked is obtained in real time in the world coordinate system.
Author supplied keywords
Cite
CITATION STYLE
Zhongmin, W., & Shanjia, D. (2018). Mean-shift algorithm for 3D spatial positioning. In Procedia Computer Science (Vol. 131, pp. 446–453). Elsevier B.V. https://doi.org/10.1016/j.procs.2018.04.251
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.