A camera-equipped sprayer will scan wheat fields in real time, identify patches of black-grass, and apply herbicide only where needed. Black-grass is the UK’s most destructive weed, costing cereal growers millions in lost yield each year. Overuse of herbicides has driven the evolution of resistant strains, while blanket spraying damages beneficial species and pushes costs beyond what many farmers can afford. Current methods cannot map weed distribution fast enough or precisely enough to allow targeted spraying early in the season. This project builds a self-propelled sprayer that uses artificial intelligence to distinguish black-grass from crops, generates precise weed maps, and controls each nozzle individually. The data will feed into agronomic models on the xarvio platform to produce bespoke recommendations for herbicide type and variable application rates. If successful, the system could cut herbicide use dramatically while maintaining control efficacy, lowering costs for growers and reducing environmental harm from off-target effects. The field data will also support further research into resistance monitoring and sustainable weed management.
View original technical description
Black-grass (_Alopecurus myosuroides_) is the UK's most pernicious weed, causing considerable yield losses each year and threatening the sustainability of UK cereal production. Herbicides remain a key component of tactics to control this species, yet evolution of resistance due to overuse threatens their efficacy. Environmental considerations also raise concerns about such widespread herbicide use, including the potential for damaging off-target effects on potentially beneficial species. Finally, the cost to growers of existing herbicidal strategies are becoming prohibitively expensive. Precision spraying of herbicides specifically to weed patches, rather than the whole field, offers the potential to help mitigate these issues, considerably reducing herbicide use whilst maintaining effective control of the weed. To facilitate this, there is a need to develop systems for accurately and rapidly mapping the spatial distribution of the weed within the field, at time-points early enough to enable intervention. Spot application needs to be robust enough to avoid missing weeds and provide sufficient confidence for growers to support their uptake. And finally, software and machinery need to be in place for delivering the targeted, precision herbicide application. This project will enable UK SME Chafer to use the Bosch/BASF 'smart sprayer' concept and technology to develop, build and evaluate a camera-equipped self-propelled sprayer for black-grass mapping and precision patch- or spot-spraying. The artificial intelligence to identify weeds and the ability to generate precise weed maps will allow Rothamsted and BASF to conduct further research. The outcome of the project will allow field data generation to be used for agronomic recommendations. The data will be analysed on the xarvio(tm) platform, and models developed by Rothamsted will be used to develop bespoke recommendations of herbicide choice and variable-rate application of pre- and post-emergence herbicides. These plans will be implemented using precise (individual nozzle) control of herbicide application by the sprayer system, with performance evaluated against conventional 'whole-field' management in herbicide usage, cost reductions, and control efficacy.
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know