Design of Visual Navigation System of Farmland Tracked Robot Based on Raspberry Pie

With the rapid development of Internet of Things Technology, agricultural robot technology has made rapid progress. A visual navigation system of farmland tracked robot based on and Raspberry Pi and monocular camera are designed. The graying factor to realize gray image, median filtering, Otsu threshold segmentation and morphological processing to segment graying image from green plants and background are analyzed and implemented. A convenient and relatively accurate central path extraction algorithm and selection method of region of interest are proposed. Also the calculation of navigation parameters is proposed to control the forward and steering operations of the robot. The design has good reliability, strong anti-interference ability, low cost and wide application range agricultural production.