Aerial Drone Mapping and Trajectories Generator for Agricultural Ground Robots

Abstract

Precision agriculture concept using robotics technology has attracted much interest. The concept requires increasing crop yield and quality while reducing labor costs by better cultivating the farmland using robotics technology. The first step of such technology requires sophisticated maps to get better accuracy. Conventional maps, such as scale maps and topological maps, are susceptible to noise and might not represent the real terrain of farmland, while digital terrain models acquired from satellites and other advanced acquisition devices are quite costly. In this paper, a system of an automatic trajectories generator for the paths of ground robots is developed. Aerial images taken from drones are employed to reconstruct a map that better representing the real terrain. A contour detector is then developed to get the farm boundaries, which are then followed by generating trajectories to be sent to the ground robots. We believe that this work will bring many benefits to agriculture.