A Agricultural Spraying and Fertilization Robot based on Visual Navigation

Abstract

The research and popularization of agricultural intelligence and automation is the guarantee of modern agriculture. At present, Autonomous navigation of agricultural fertilization robot can effectively reduce the use of labor, and greatly improve the work efficiency, which has become the focus point of agricultural robot research. In this paper, a agricultural fertilization robot with autonomous navigation was studied, also a roller piercing fertilizer applicator was designed and developed, which can avoid the waste of surface fertilization, and improve fertilizer efficiency. The robot mainly includes the control of chassis, farmland image processing, navigation path planning and mechanical design of spraying and fertilization. For robot control, the Cortex-m4 is selected to control the motors of chassis. For farmland image processing, through the color space selection, filtering, enhancement, and segmentation, the useful information can be obtained for visual navigation. The Raspberry Pi is employed to process the image information. Through serial communication, Raspberry Pi can send instructions to Cortex-m4 for controlling robot chassis and receive the feedback. Because the Raspberry Pi takes a long time to completely process the image, the image processing navigation algorithm is improved. By comparison, the processing speed has been increased by four times, which effectively guarantees the real-time of robot autonomous navigation. The experiments results show that the robot is suitable for arid areas and can be used for deep fertilization, the improved algorithm can effectively enable the robot to carry out for autonomous navigation, fertilizing and spraying.