An Autonomous Guided Field Inspection Vehicle for 3D Woody Crops Monitoring

Abstract

This paper presents a novel approach for crop monitoring and 3D reconstruction. A mobile platform, based on a commercial electric vehicle, was developed and equipped with different on-board sensors for crop monitoring. Acceleration, braking and steering systems of the vehicle were automatized. Fuzzy control systems were implemented to achieve autonomous navigation. A low-cost RGB-D sensor, Microsoft Kinect v2 sensor, and a reflex camera were installed on-board the platform for creation of 3D crop maps. The modelling of the field was fully automatic based on algorithms for 3D reconstructions of large areas, such as a complete row crop. Important information can be estimated from a 3D model of the crop, such as the canopy volume. For that goal, the alpha-shape algorithm was proposed. The on-going developments presented in this paper arise as a promising tool to achieve better crop management increasing crop profitability while reducing agrochemical inputs and environmental impact.