4.7 Article

Inter-row vision guidance for mechanical weed control in sugar beet

Journal

COMPUTERS AND ELECTRONICS IN AGRICULTURE
Volume 33, Issue 3, Pages 163-177

Publisher

ELSEVIER SCI LTD
DOI: 10.1016/S0168-1699(02)00005-4

Keywords

mechanical weed control; vision guidance; inter-row guidance

Ask authors/readers for more resources

Despite a general move over the last 20 years towards low volume overall spraying techniques, mechanical weeding using inter-row hoes is still practised by a significant number of UK sugar beet growers. Reasons for this include the control of difficult weeds e.g. volunteer potatoes, thistles and larger weeds surviving herbicide application, as well as the control of weed beet. However, for general weed control, savings in herbicide cost have to be balanced against a relatively low hoeing work rate. Increased use of inter-row hoeing to achieve the potential economic and environmental benefits, can only be accomplished by the introduction of high workrate equipment. This paper describes research, building on earlier work conducted in cereals, which addresses the problems of automatically guiding hoes between rows of sugar beet at high speed. Results indicate that the previously used combination of Kalman filter row tracking and bandpass filter row location provide a sound basis for following rows of larger sugar beet plants. Smaller plants could also be followed if the single pixel scan lines used by the bandpass filter algorithm were replaced by bands formed by merging columns of pixels to increase the information content. The hoe was evaluated at 6 kph under a variety crop conditions ranging from two true leaves with a population of 1700/m(2) newly germinated weeds, through to a relatively clean crop in which crop plants had just started to touch in-the-row. Lighting conditions included diffuse sky and direct sun illumination in which the tractor and hoe cast shadows in the image region. Under this range of conditions standard deviation in hoe lateral error relative to crop rows remained within 16 mm and mean bias never exceeded +/- 10 mm. Additional tests to evaluate the hoe's ability to cross gaps in which all four of the crop rows used for guidance had been removed showed that error did not become unacceptable until gap length exceeded 4 in. (C) 2002 Elsevier Science B.V. All rights reserved.

Authors

I am an author on this paper
Click your name to claim this paper and add it to your profile.

Reviews

Primary Rating

4.7
Not enough ratings

Secondary Ratings

Novelty
-
Significance
-
Scientific rigor
-
Rate this paper

Recommended

No Data Available
No Data Available