Model Based Tracking for Navigation and Segmentation

Abstract

An autonomous vehicle has been developed for precision application of treatment on outdoor crops. This document details a new vision algorithm to aid navigation and crop/weed discrimination being developed for this machine. The algorithm tracks a model of the crop planting pattern through an image sequence using an extended Kalman filter. A parallel update scheme is used to provide not only navigation information for the vehicle controller but also estimates of plant position for the treatment system. The algorithm supersedes a previous Hough transform tracking technique currently used on the vehicle which provides navigation information alone, from the rows of plants. The crop planting model is introduced and the tracking system developed, along with a method for automatically starting the algorithm. In applications such as this, where the vehicle traverses unsurfaced outdoor terrain, “ground truth” data for the path taken by the vehicle is unavailable; lacking this veridical information, the algorithm's performance is evaluated with respect to human assessment and the previous row-only tracking algorithm, and found to offer improvements over the previous technique.