Root zone fertilization robot based on YOLOv5 stalk identification

Xia Li; Mengchao Hu; Fangtao Duan; Tengfei Tang; Xiwang Du; Zhangjun Jiang

doi:10.21162/PAKJAS/24.86

Xia Li Tianjin Key Laboratory for Advanced Mechatronic System Design and Intelligent Control, School of Mechanical Engineering, Tianjin University of Technology, Tianjin 300384, China; National Demonstration Center for Experimental Mechanical and Electrical Engineering Education, Tianjin University of Technology, Tianjin 300384, China

Mengchao Hu Tianjin Key Laboratory for Advanced Mechatronic System Design and Intelligent Control, School of Mechanical Engineering, Tianjin University of Technology, Tianjin 300384, China; National Demonstration Center for Experimental Mechanical and Electrical Engineering Education, Tianjin University of Technology, Tianjin 300384, China

Fangtao Duan Tianjin Key Laboratory for Advanced Mechatronic System Design and Intelligent Control, School of Mechanical Engineering, Tianjin University of Technology, Tianjin 300384, China; National Demonstration Center for Experimental Mechanical and Electrical Engineering Education, Tianjin University of Technology, Tianjin 300384, China

Tengfei Tang Tianjin Key Laboratory for Advanced Mechatronic System Design and Intelligent Control, School of Mechanical Engineering, Tianjin University of Technology, Tianjin 300384, China; National Demonstration Center for Experimental Mechanical and Electrical Engineering Education, Tianjin University of Technology, Tianjin 300384, China

Xiwang Du Tianjin Key Laboratory for Advanced Mechatronic System Design and Intelligent Control, School of Mechanical Engineering, Tianjin University of Technology, Tianjin 300384, China; National Demonstration Center for Experimental Mechanical and Electrical Engineering Education, Tianjin University of Technology, Tianjin 300384, China

Zhangjun Jiang Tianjin Key Laboratory for Advanced Mechatronic System Design and Intelligent Control, School of Mechanical Engineering, Tianjin University of Technology, Tianjin 300384, China; National Demonstration Center for Experimental Mechanical and Electrical Engineering Education, Tianjin University of Technology, Tianjin 300384, China

The precise amount of liquid fertilizer and the suitable location for fertilization during the growth of crops will promote crop growth and improve fruits quality. However, existing fertilizer application methods, while improving fertilizer utilization to some extent, can have adverse effects such as soil salinization and weed breeding. To this end, this paper developed a robot platform for root zone fertilization based on stalk identification, used the D435i depth camera and YOLOv5 detection algorithm to identify single seedling stalks and positioning the compound point of the root and stalk, which provides corresponding guidance for the movement of the fertilization robotic arm. At the same time, the fertilization pressurization system quickly delivers the liquid material in the fertilizer tank carried by the robot to the soil near the stalk of crops to achieve the purpose of reducing fertilizer and increasing its effect. The experimental results show that the average identification accuracy of the improved target detection algorithm reaches 99.3%. Under the premise that the distance between the fertilization point and the seedling stalk is specified as 30 mm, the average absolute error between the end of the fertilization needle and the center point of the stalk bottom is 2.15 mm, and the maximum error distance is 3.9 mm. This method can identify the stalk and locate the center point of the stalk bottom for single corn seedlings, and simultaneously complete the liquid fertilizer application at specific locations.