Autonomous interrow hoeing robot based on image processing

Today, mobile robot technology is widely used in many sectors. Mobile robots are not physically fixed to a specific point, they can move in a predefined area and perform the desired tasks. Thanks to these robots, repetitive operational operations that people will do can be done quickly and independently of humans. In this thesis study, an easy-to-use, farmer-friendly, fully autonomous agricultural robot that prevents high yield losses caused by weeds in the production of tobacco plants, which are among the important export products of our country and provide yield increase by aerating the soil was designed. First of all, research was conducted on the area where the robot will work and the effects of the loads coming from the soil to the robot were calculated. As a result of the calculations, mechanical and electronic designs were completed. Mechanical analyzes were performed in the simulation and motion analysis environment of the SolidWorks program. Then, image processing software was developed that allows the robot to move autonomously between the rows. The image processing software was developed using the deep learning algorithm YOLO (You Only Look Once). As a result, the design of a mobile agricultural robot that will provide autonomous weed control in tobacco farming was realized with this thesis study.