Vision-guided spraying technology targets problem weeds

IR-4 researchers explore how, by ″seeing″ weeds, targeted applications can reduce crop injury, increase control, and manage suckers on grapes.

What to apply, when to apply, and how to apply are key considerations when it comes to herbicides—particularly for weed control on perennial crops. Two IR-4 research collaborators are working on the ″how″ part with a brand-new precision-spraying technology. Lynn Sosnoskie, Assistant Professor of Weed Ecology and Management in specialty crops at Cornell University, and Thierry Besançon, Associate Professor and Extension Weed Specialist for specialty crops at Rutgers University, have been searching for new herbicides to use safely on perennial crops, with a focus on increasing efficacy, managing crop injury, and reducing the total amounts of product applied.

The trend toward reducing the use of herbicides and other pesticides due to various factors—including environmental concerns, the development of resistance, and regulatory pressures—led Besançon and Sosnoskie to investigate precision spraying, which ultimately uses less herbicide in the field. The project was funded under IR-4’s Integrated Solutions platform.

″It was Lynn’s idea to use a vision-guided sprayer that can detect the green color—the chlorophyll—of the weed,″ Besançon says. ″So instead of doing a broadcast or banded application where you’re covering the entire surface of the soil with herbicide, you’re just targeting the weeds. It has two benefits: potentially reducing the amount of herbicide we are using, and possibly allowing us to use herbicides not currently listed for that crop since we can use it more safely (because we are only targeting the weeds).″

WEED-IT sensors can detect weeds by sight (sensing the green of chlorophyll), enabling precise, targeted application of herbicides.

Putting Precision Sprayers to the Test

An ATV carrying a sprayer fitted with a WEED-IT sensor was driven through the vineyard or field, only spraying when weeds triggered the sensors. For the IR-4 field trials, Besançon focused on weed control under the canopies of blueberries while Sosnoskie worked on weed control under the canopies of grapes, as well as the management of grape suckers. Besançon and Sosnoskie applied up to eight herbicides at different rates via continuous banded application underneath the canopy or by the vision spray technology. Data collected included crop injury, yield, weed cover, weed control, and the amount of product used. 

Sosnoskie says the vision sprayer technology worked quite well in grapes; the high canopy made it easy for the sensors to identify weeds and suckers quickly. ″Early in the season, the only green you see under the canopy is the weeds and the grape suckers, so there isn’t much risk of crop injury,″ she says. ″And reducing suckers with this method saves a lot of manual labor, which of course is important to growers.″

Besançon says the results were mixed with blueberries due to their growth habit. ″Blueberries send up new stems from the crown, which are needed to keep the plants productive,″ he says. ″So we had more crop injury because some small, new shoots were also getting hit. But I think part of the problem is we had too many weeds, and it was constantly spraying, which increased the risk of injury to the crop. I think this technology has potential, but I’d like to gather more data before giving it a full recommendation on blueberries.″

Both researchers say this technology may be more effective under low weed pressure. It’s best to use it early in the season, after a postharvest burndown and when weeds first emerge in the spring. It’s not a replacement for a traditional weed program—it is an additional tool for managing problem weeds that have escaped a spring residual spray or that have germinated later in the season. 

″It’s meant for growers with a good handle on weed control. But this technology could increase their success by another 10% or 20% and target the problematic weeds that escape,″ Sosnoskie says.

The team’s paper on their highbush blueberry work, co-authored by Besançon, Sosnoskie, and Roger Batts, IR-4’s principal weed scientist, has been published by HortTechnology.

Technology May Open Up Possibilities for More Herbicide Tools

A few herbicides included in the trials are not currently labeled for blueberries or grapes, so the trials contributed valuable data to IR-4’s registration efforts. ″We used fluroxypyr (Starane^® Ultra from Corteva Agriscience), which is already in the IR-4 system for testing on a few crops, so we chose it to provide data for possibly registering this product on these crops,″ Besançon says. ″Lynn [Sosnoskie] used mesotrione and 2,4-D because those aren’t listed for grapes.″

″Group 4 herbicides (including 2,4-D) and grapes don’t mix well, but some of those tools are the best ones we have for management of some of our weed species,″ Sosnoskie says. ″So this data will also help us develop an IR-4 Integrated Solutions trial looking at various [auxinic] products with different types of application technologies [including precision spray technology] to see if we can improve the safety of these tools on crops that they were not able to be used on before.″ 

Vision spraying technology could even make listed herbicides more usable. Besançon cites yellow nutsedge as a problem weed with only one (expensive) herbicide registered for controlling it in blueberries. ″People are reluctant to use it because it’s so costly, especially if you do a broadcast application,″ he says. ″Using precision technology would allow growers to use this efficiently and significantly reduce costs. Sometimes farmers send a crew out to manually spot spray, but the labor is also expensive.″

In this project, Sosnoskie found that the vision sprayer technology worked quite well in grapes; the high canopy made it easy for the sensors to identify weeds and grape suckers quickly.

IR-4’s Integrated Solutions Platform Helps Growers with Sustainability Goals

In 2018, IR-4 launched the Integrated Solutions platform to evaluate emerging technologies, biopesticides, and other novel tools in combination with conventional pesticides to find solutions to the complex problems growers face. It fits well with the U.S. Environmental Protection Agency’s herbicide strategy, which aims to protect over 900 endangered species. By finding solutions that provide growers with ecologically sound tools that still control pests and weeds effectively, the Integrated Solutions platform will aid both the EPA’s goals and help growers meet the new regulations. In particular, precision spraying is mentioned in the EPA’s strategy ″as a tool for reducing the proportion of the field treated (and reducing non-target exposure from runoff or erosion).″

New tools and strategies cannot come too soon. ″With climate change, perennial weeds are expected to become much more of a problem, and they are already some of the weeds we are having a hard time managing,″ Sosnoskie says. ″We will also see more pressure on the beneficial organisms and species we want to protect, so precision spraying is just another tool to help mitigate the effects of herbicide products on the environment.″

Next Steps Towards Implementation

This initial trial is a springboard for more research. Next, Sosnoskie and Besançon plan to equip a Farm-ng AMIGA autonomous platform with WEED-IT sensors to increase labor-saving. Other than the person who fills the sprayer and sends it on its way, no human labor will be required. They have already worked with this technology and see significant potential for growers. Sosnoskie says that when she demonstrated the vision spraying technology and talked about the results she was seeing, growers were optimistic. 

The opportunity for further automation using robots like the Amiga also sparks interest. When their team acquired the fleet of equipment used by Sosnoskie and Besançon, it cost less than $40,000 (around $15,000 for one system and around $25,000 for the other). Sosnoskie noted that growers may explore grant funds to pay for this equipment since it is a sustainable solution.

Besançon agrees, emphasizing that while it’s always great to have new technology, it has to be financially realistic for those using it. ″It’s not too expensive,″ he says. ″The vision sprayer can be put on a tractor or an ATV. It’s pretty easy to adjust and adapt the system, so you don’t need an expert to come in every time you use it. So I think this can translate relatively quickly to growers.″