Background Reducing pesticide use in specialty crops is a major societal and challenge. Although high-end sprayers can achieve up to 25% plant protection product (PPP) savings through detailed calibration, the EU’s Farm-to-Fork strategy targets a 50%. Many simple tree and bush crop sprayers remain inefficient, with up to half of the applied PPP lost to drift, run-off, and ground deposition. However, advanced sensing technologies combined with precision spraying systems, offer opportunities for substantial additional PPP savings. Methods Field trials were conducted in commercial citrus, almond, and olive orchards across Spain. A Fede Smartomizer H3O sprayer was used as baseline. Further, a real time spot spraying system - comprising 3D stereo vision cameras and individual nozzle control - was mounted to collect real-time canopy geometry data. This information was transmitted to a spot spraying enabled sprayer with independently actuated nozzles. Nozzle activity was logged. Spray deposition was evaluated using hydrosensitive paper. Savings of the spot spraying system, commercially referred to as “AIs”, were computed relative to the constant rate baseline application. Results Across all test sites, AIs-enabled spot spraying significantly reduced PPP use without compromising canopy coverage. Spray volume savings were 43% in citrus, 48% in almond, 56% in intensive olive, and 67% in extensive olive orchards. Moreover, mathematical closed-form expressions are derived to indicate the potential savings associated with the application of spot spraying. Conclusion Concluding, real time spot spraying in tree and bush crops, like implemented thought the AIs system, enables substantial additional PPP savings beyond what is achievable with well-calibrated constant-rate sprayers. Savings depend on canopy variability and the relationship between canopy structure and sprayer resolution. Closed-form upper bounds are derived indicating savings potential of an ideal spot sprayer with infinite resolution.
Berger et al. (2026) studied this question.