Autonomous Vehicle for Paddy Field Weeding by Image Feedback Control

Abstract

This paper presents a control method so that an agricultural vehicle for paddy field weeding runs adaptively in its environment. We have realized that the vehicle autonomously tracks a row of rice and turns at a headland in paddy field by image feedback control. In order to extract the rice plant and to recognize the headland, without the influence of the weather, floating matters in the paddy field and growth of the rice plant, we use a color ingredient extraction method using RGB values of a color image. Moreover, a tracking line is detected by Hough transform to the extracted image. A gyro and an encoder enable the vehicle to turn the headland.