Design and implementation of a gardening mobile robot with embedded face-tracking system

This paper proposes a gardening robot capable of visual tracking and obstacle avoidance. Firstly, YCbCr color space transformation is utilized for skin detection. Then, dilation/erosion/closing and 8-neighbors methods are adopted to search the maximum area of desired object and eliminate noise. Finally, center coordinate of object is acquired. The tracking object can be set beforehand or immediate identification. The micro-controller is employed to receive coordinate information and conduct steering control. The infrared sensors module is used to avoid obstacle. The proposed system designs one human-machine interface at remote end for easy surveillance and emergent control of robot location. This system is implemented in one prototype of gardening robot, which can share farmers' task and provide carriage of goods.