Vegetable growers often use herbicides to control weeds and prevent crop loss. However, researchers at the University of Florida are working on ways to reduce herbicide use through the application of artificial intelligence (AI).

Weeds are a significant issue for vegetable farmers, competing with crops for resources like water, nutrients, light, and space. As a result, herbicides are used in approximately 90% of vegetable production in the United States.

Nathan Boyd, a UF/IFAS weed scientist and professor at the Gulf Coast Research and Education Center (GCREC), explained, "We are developing multiple AI-powered herbicide application units, each with a different approach." These AI systems aim to apply herbicides more efficiently by targeting specific areas where weeds grow.

In a recent study, Boyd and his team developed and tested a precision-spraying system designed to apply herbicide directly to the soil beneath holes in plastic mulch, where tomatoes are grown. This system, referred to by Boyd as an “AI smart sprayer,” was able to locate the holes 86% of the time, reducing herbicide use by over 90% in tomato fields at the research center.

By targeting only the holes in the mulch, where weeds can emerge, the system applies herbicide only to the areas where it is needed. 

"We are focusing on applying herbicides only through the holes in the plastic mulch, rather than across the entire bed," Boyd explained. "This results in a significant reduction in herbicide use."

Plastic mulch is commonly used in vegetable production, covering the soil where crops are grown. Traditionally, herbicides are applied to the entire soil bed before the plastic is placed, but this method is inefficient and can lead to unnecessary chemical use, higher costs, and potential crop damage.

Boyd and his team have been exploring the use of AI for both pre-and post-emergence herbicide applications for several years. In 2018, Boyd collaborated with Arnold Schumann, a professor at the UF/IFAS Citrus Research and Education Center, to develop an early version of the precision sprayer.

However, that version did not incorporate AI to improve the precision of identifying the holes in the mulch. The new technology is able to find the holes more accurately. The next steps involve making this technology available to farmers and assessing its broader impact. 

"Further research is necessary to understand the economic implications of reduced herbicide use and evaluate the cost-effectiveness of this approach," Boyd said.

Related articles: Agroview: The UF/IFAS AI multi-problem solution