Online Extrinsic Calibration of RGB and ToF Cameras for Extraterrestrial Exploration

Extraterrestrial exploration is one of the most significant missions for the development of countries. Different sensors are used in an extraterrestrial exploration mission for different kinds of tasks, such as navigation and grasping. Correct extrinsic parameters for multiple sensors is fundamental for the performance of high-level tasks. Although the extrinsics can be accurately calibrated offline, they easily become inaccurate due to the vibration of the vehicle during the tasks. In this paper, an online extrinsic calibration algorithm is proposed for extraterrestrial exploration. Extrinsic parameters of RGB and ToF (time-of-flight) cameras are automatically adjusted while navigating the environment. The online calibration is achieved by the alignment of edges extracted from the RGB image and ToF intensity image, respectively. An edge registration algorithm is proposed and the probabilistic properties of pixels on edges are exploited for the alignment of the edges extracted from heterogeneous sensors. Experiments executed in the simulated extraterrestrial environment demonstrate effectiveness and efficiency of the proposed algorithm.