Reducing herbicide use in row crops with targeted application methods treating detected weeds in small patches or spots

Previous work (described in the related project) led to the development of an image analysis-based weed detection system linked to a spot spray control mechanism. The system was initially developed around the specific problem of treating volunteer potatoes within onion, carrot and parsnip crops.  Discrimination of live plant material from background was on the basis of colour and a number of criteria were used to determine if plant material was crop or weed.  

This follow-on project aimed to extend the approach developed in the earlier work to:

• Enable a wider range of formulation types to be used;
• Address a wider range of target weed species in a wider range of crops particularly onions and leeks;
• Have the ability to treat patches of weeds as well as using spot applications directed at single weeds.

A practical system was developed for treating large weeds in vegetable and sugar beet crops by the spot application of selective and non-selective herbicides and to detect and treat patches of weeds in widely spaced row crops.  Field trials with a full-scale experimental machine showed that the system was capable of:

• Detecting high percentages (circa 95%) of large weeds in crops such as onion, leek and sugar beet and applying selective and non-selective herbicides to detected weeds to give high levels of weed control.
• The application of non-selective herbicides at the full recommended field dose gave a rapid and complete kill of detected weeds with commercially acceptable levels of crop damage – this kill was noticeably faster and more efficacious than when using selective herbicide mixtures.
• Measurements of the spray deposits on crop plants adjacent to spot treated weeds showed that the target weeds had at least an order of magnitude greater deposits than did the crop plants. This was also reflected in assessments of glyphosate residue levels in crop plants growing very close to spot treated weeds where no detectable residues were found in crop plants of a harvestable size.
• A system of detecting patches of small weeds in row crops based on an assessment of plant material in the inter-row gap was shown to be feasible and the basis for the targeted application of selective herbicide treatments over the crop row with the option of using non- selective herbicides between the rows.
• Assessments of the economic viability of operating a spot application system as part of an intelligent band sprayer concluded that this would give financial savings in many horticultural crops and that the potential for such savings will increase as the availability of selective herbicides reduces.