Integrating computer vision and precision sprayers for targeted green fruit chemical thinning

Abstract Chemical thinning is a vital practice in apple production, used to manage crop load, improve fruit quality, and ensure consistent annual yields. This study investigated the refinement of chemical thinning through integration of a commercial computer vision system with a precision variable-rate sprayer (VRA), aiming to reduce chemical usage while maintaining efficacy. Field trials were conducted in a mature ‘Fuji’ apple orchard, comparing three treatments: (1) precision approach using computer vision-generated fruitlet maps and a VRA sprayer; (2) conventional approach with a uniform spray rate; and (3) an untreated control. A Vivid XV3 imaging platform mounted on a utility vehicle was used to detect and count fruitlets per tree, generating a georeferenced prescription map. This map was uploaded to the Intelligent Spray Application ® system, which modulated chemical flow in real time based on GNSS-guided positioning. Thinning effects were evaluated using follow-up scans, manual fruitlet counts, and harvest data, including total yield and average fruit mass. Both the precision and conventional approaches achieved comparable reductions in fruit density and similar fruit size. Notably, the precision treatment used approximately 18% less chemical thinning agent. These findings underscore the practical advantages of integrating computer vision and variable-rate application technologies for more efficient and sustainable orchard management. By targeting chemical inputs based on tree-specific conditions, this approach offers growers a promising tool to optimize input use, reduce environmental impact, and maintain or enhance fruit production outcomes.