Increasing Work Rate in Vision Guided Precision Banded Operations

Concerns over the economic and environmental sustainability of farming are leading to an increased interest in techniques to reduce crop protection inputs by better targeting. One way to improve targeting is through physically aligning treatment devices with crop rows. Manual guidance is a difficult and stressful task due to the high concentration required to maintain acceptable tolerances, typically ±25 mm. Under normal farm conditions, global positioning system based technologies are unable to reliably offer this performance. Commercial implement guidance systems based on computer vision have demonstrated that the required accuracy can be achieved. However, working width is limited due to the need to match implements with drill or transplanter width in order to avoid misalignment at the interface between bouts.This paper describes an integrated vision-based system for tracking multiple bouts that has been tested in two example applications. The first was an inter-row hoe for use in cereals with three independently guided 4 m wide sections, each with its own camera, operating on 12 m tramlines. Trials showed that the standard deviation in the lateral position was 10 mm at 10 km h−1. The second exemplar based on a conventional boom sprayer was a precision band sprayer for vegetables spanning five 4 m wide beds. Trials indicated the standard deviation in lateral position was 22 mm at 12 km h−1.