Accurate Robot Localization using a Single Camera

C. Ricolfe-Viala and A.-J. Sánchez-Salmerón (Spain)


robot localization, camera calibration, Kalman filter.


Robot localization is an important issue in order to provide them truly autonomous capabilities. Several techniques exist. Basically they mix internal and external robot measurements using Kalman filter. In this paper a technique based on a single camera is presented. Robot localization is done using planar landmarks. Landmarks are used to calibrate the camera and to compute its position and orientation optimally in a global reference system. Camera calibration is carried out using Kalman filter updating the camera parameters with new data coming from a new image. It increases the accuracy of the localization process. Low uncertainty of robot position and orientation is achieved.

Important Links:

Go Back