BUSINESS VegNET

The OWL system
uses camera
detection and
AI algorithms
to identify
green weed
plant matter.
Open Weed Locator Demo Day showcases DIY spraying innovation in Manjimup
When 11 members of the Warren Cauliflower Improvement Group jumped on a plane to Copenhagen and the Netherlands in 2025, they weren’t just heading on a study tour, they were exploring the future of on-farm weed control.
Words Lisa Hasson, Regional Support Officer, vegetablesWA
LED by the Warren Cauliflower Improvement Group President, Jake Ryan, the group travelled to learn how to build a DIY artificial intelligence sprayer using open-source technology. Joining them was former VegNET WA Regional Development Officer Katrina Hill, supporting the initiative and its potential application back home in Western Australia’s horticulture sector.
Fast forward to 27 February 2026, and that learning came full circle, with the first local prototype demonstrated at a field day in Manjimup.
From study tour to shed-built innovation
The VegNET funded 2025 study tour focused on practical knowledge instead of theory. The goal was simple — to understand how growers could build and adapt an affordable AI-based sprayer system themselves instead of relying solely on expensive commercial platforms.
Dr Guy Coleman introduced the group to the Open Weed Locator, known as OWL, the DIY, open-source algorithm device designed to identify weeds in real time and activate targeted spraying only where needed.
For growers facing soaring chemical costs, stubborn resistance problems and increasing pressure to farm sustainably, the idea couldn’t have been more relevant.
Rather than blanket spraying entire paddocks, the OWL system uses camera detection and AI algorithms to identify green weed plant matter against green crop foliage, triggering a spray nozzle only when required.
So far, the numbers point to potential chemical savings of up to 95% in the right conditions.
What is the Open Weed Locator?
The Open Weed Locator is not a commercial black-box product, which is precisely what makes it different.
It is a DIY, open-source system that offers a significantly cheaper alternative to commercial optical sprayer options. Designed to be built and maintained by growers, it removes the barrier of ongoing subscription or licensing costs.
For horticulture growers operating on tight margins, affordability is a major advantage. While many commercial precision spraying systems require substantial capital investment and ongoing fees, OWL provides a lower-cost entry point into precision spraying technology.

The OWL system uses camera detection and AI algorithms to identify green weed plant matter.

However, it is still early days. The prototype demonstrated on 27 February 2026 was the first local build, and further development is required — particularly in the creation of a crop library before full-scale adoption can begin.
Demo Day: 27 February 2026, Manjimup
The February demonstration day marked an important milestone, being the first time the locally built OWL prototype was showcased infield to Western Australian growers.
The atmosphere on the day mixed curiosity with plenty of practical questions and a healthy dose of cautious optimism. Attendees saw firsthand how the camera detection system identified plant matter and triggered targeted spray outputs. Discussion quickly shifted to how adaptable the system could be across vegetable crops, what was required to build one, how long calibration might take and how it would perform under variable field conditions.
Importantly, the demo was not framed as a commercial release. It was presented in its current form as a working prototype and a base to build on. The message across the day was that this is grower led innovation shaped by local challenges and practical realities.
What happens next?
While the technology shows significant promise, the next major step is developing a robust crop library.
For the OWL system to effectively differentiate between crop plants and weeds, especially in horticulture systems where crop spacing and canopy structures vary, it requires trained datasets specific to local crops and growing conditions. This means collecting images across different growth stages, testing the system across multiple crop types and refining the algorithm for Western Australian environments.
To help growers dig deeper, a set of supporting webinars and a podcast have already been released. Together they walk through the study tour and the build, giving growers a clearer sense of what’s involved and what opportunities this technology could open up.
Why this matters for WA horticulture
Rising input costs, labour constraints and chemical resistance pressures continue to challenge vegetable growers across the state.
The Open Weed Locator project highlights how growers can lead the charge on new technology instead of relying on commercial systems that often arrive late and out of financial reach.
The Open Weed Locator project is practical, collaborative and still evolving.
The February demo demonstrated that AI spraying does not have to be inaccessible for horticulture businesses. However, continued testing, data development and grower involvement will be critical in determining how quickly and effectively the system can be integrated into commercial vegetable production.
For those who were unable to attend the field day, the event highlights, webinars and podcast offer an in-depth look at what has been achieved so far, and where the project is heading next.
The focus has moved beyond debating whether AI fits horticulture and toward how growers can make it work for their own systems.
MORE INFORMATION
Contact Lisa Hasson, Regional Support Officer, on 0459 952 297 or lisa.hasson@vegetableswa.com.au