An automatic calibration between an omni-directional camera and a laser rangefinder for dynamic scenes reconstruction

摘要

An automatic approach for the calibration of an omni-directional camera with a laser rangefinder(LRF) on a mobile robot is presented. We improve the current calibration approach to make it adjust to the omni-directional camera with LRF. Then we focus on building the static-map for the dynamic scene using a mobile robot equipped our camera-LRF system. In this work, we propose a novel approach for updating 3D color map with a Guassian mixture model. Dynamic obstacles are detected combined with 3D and color feature using this model. In the experiment, on one hand, the robustness of our calibration approach is tested by the reconstruction of the omni-directional scenes with color and 3D feature. On the other hand, our map model for dynamic scenes was tested on our system and KITTI dataset, the result shows our approach can reconstruct static map and remove almost dynamic obstacles.