Simulation-based nozzle density optimization for maximized efficacy of a machine-vision weed control system for applications in turfgrass settings

Abstract

Targeted spray application technologies have the capacity to drastically reduce herbicide inputs, but to be successful, the performance of both machine vision (MV) based weed detection and actuator efficiency need to be optimized. This study assessed 1) the performance of spotted spurge recognition in ‘Latitude 36’ bermudagrass turf canopy using the You Only Look Once (YOLOv3) real-time multi-object detection algorithm, and 2) the impact of various nozzle densities on model efficiency and projected herbicide reduction under simulated conditions. The YOLOv3 model was trained and validated with a dataset of 1,191 images. The simulation design consisted of 4 grid matrix regimes (3 × 3, 6 × 6, 12 × 12, and 24 × 24), which would then correspond to 3, 6, 12, and 24 non-overlapping nozzles, respectively, covering a 50-cm wide band. Simulated efficiency testing was conducted using 50 images containing predictions (labels) generated with the trained YOLO model and, by applying each of the grid matrixes to individual images. The model resulted in prediction accuracy of a F1 Score of 0.62 precision of 0.65 and recall value of 0.60. Increased nozzle density (from 3 to 12) improved actuator precision and predicted herbicide-use efficiency with a reduction in false hits ratio from ∼30% to 5%. The area required to ensure herbicide deposition to all spotted spurge detected within images was reduced to 18% resulting in ∼80% herbicide savings compared to broadcast application. Slightly greater precision was predicted with 24 nozzles, but not statistically different from the 12-nozzle scenario. Using this turf/weed model as a basis, optimal actuator efficacy and herbicide savings would occur by increasing nozzle density from one to 12 nozzles with the context of a single band.