Binocular camera-based a docking system for an amphibious spherical robot

In order to fulfill the task about autonomous recovery and charging of robots, a set of robotic docking system based on binocular stereo vision was initially designed and implemented. A binocular camera was used in our docking system. Through the image captured by binocular camera of the surrounding environment, we can obtain the two kinds of image, which include color image and depth image. And in order to get accurate depth information, we conducted a binocular camera calibration experiment. Then according to the depth information, the robot recognize the docking receptacle also as a marker fixed in the pool, then using KCF (Kernelized Correlation Filters) algorithm to track the target and making robot arrive the docking destination accurately. The docking system can be implemented by the process above. During the experiment, we selected the receptacle in the image, the robot recognized target and arrived at the designated location. The experimental results demonstrated the veracity and robustness of the proposed docking system based on binocular stereo vision.