New algorithms extend targeted treatment to additional crops and weeds, with 55 further models available for user testing.
A precision spraying platform is expanding its plant-recognition portfolio for 2027, with 10 new commercial models planned and another 55 available for growers to test as development continues.
Swiss company Ecorobotix says the new models will be available commercially from March 2027 and include applications for garlic, leek and potato. Models under development include pepper, courgette and yellow nutsedge.
The additions will take the company’s ARA system beyond 120 recognition and treatment models. Its Plant-by-Plant technology uses plant identification to direct applications towards individual targets rather than spraying the whole field uniformly.
Ecorobotix reports that crop models reaching commercial status during 2026 were tested with customers across more than 11,000ha in over 20 countries. The company says the models achieved an average input reduction of 86% compared with broadcast spraying.
That figure is the company’s reported average across the tested models, rather than a guaranteed saving for individual crops or applications. Performance will depend on the crop, target weeds, field conditions and treatment programme.
Recognition algorithm targets yellow nutsedge
One of the forthcoming models is designed to identify yellow nutsedge, an invasive weed that can be difficult to control using conventional approaches. Ecorobotix cites Agroscope research indicating yield losses of 40–71% in affected Swiss sugar beet and potato fields.
The new algorithm is intended to identify the weed and direct a non-selective herbicide onto it while avoiding the crop. Ecorobotix describes this as the first precision spraying solution capable of detecting yellow nutsedge and applying non-selective herbicide directly to the target without damaging the crop.
The model is scheduled to be available for free testing by ARA users from March 2027. Wider field experience will be important in assessing recognition accuracy and treatment performance across different weed growth stages, crop canopies and operating conditions.
Potato model tested against broadcast spraying
The potato model provides an example of the company’s development process. In a Swiss trial on potatoes at approximately 10cm high, targeted treatment was compared with conventional broadcast application, with weed-control efficacy assessed 20 days after treatment.
The results published by Ecorobotix show comparisons involving annual mercury, redshank and black nightshade. The company says the algorithm successfully controlled common weeds found in the potato crop context.
Grower feedback has also contributed to development. Belgian potato grower Luc Détroz used the model across 300ha of potatoes, reporting control of goosefoot and knotweed in a single pass, with the crop remaining weed-free until the end of the season.
The release does not provide a full trial protocol or statistical analysis, so the results should be viewed as company-reported evidence rather than an independent assessment of performance across all conditions.
Expanding the model portfolio
The new models form part of a wider move towards crop-specific recognition and treatment. The current portfolio includes applications for different crops and weed targets, with further models available for testing before commercial release.
Ecorobotix reports that more than 1,000 ARA machines are operating worldwide and that over 600,000ha have been treated. Its existing carrot model has also seen the area treated increase by more than 80%, according to the company.
For precision farming, the development highlights the importance of crop-recognition algorithms alongside the sprayer hardware. The potential to reduce application volumes depends on reliably distinguishing the intended target from the crop and delivering treatment accurately at operating speed.
The next stage will be to establish how consistently the new models perform across commercial farms, particularly for difficult weeds where conventional control options are limited.
Further information is available from Ecorobotix.
Suggested SEO details
- SEO title: Precision spraying gains 10 new plant-recognition models
- Meta description: Ecorobotix plans 10 new commercial crop models for 2027, expanding plant recognition for targeted spraying and testing 55 further algorithms.
- URL slug: precision-spraying-new-plant-recognition-models
- Primary keyword: precision spraying
- Secondary keywords: plant recognition, precision crop protection, targeted spraying, weed detection, crop-recognition algorithms, variable application
related news:
