Development and Evaluation of an Optical Sensing System for Detection of Herbicide Spray Droplets

Abstract
Real time monitoring of herbicide spray droplet drift is important for crop production management and environmental protection. Existing spray droplet drift detection methods, such as water-sensitive paper and tracers of fluorescence and Rubidium chloride, are time-consuming and laborious, and the accuracies are not high in general. Also, the tracer methods indirectly quantify the spray deposition from the concentration of the tracer and may change the drift characteristics of the sprayed herbicides. In this study, a new optical sensor system was developed to directly detect the spray droplets without the need to add any tracer in the spray liquid. The system was prototyped using a single broadband programmable LED light source and a near infrared sensor containing 6 broadband spectral detectors at 610, 680, 730, 760, 810, and 860 nm to build a detection system for monitoring and analysis of herbicide spray droplet drift. A rotatory structure driven by a stepper motor in the system was created to shift the droplet capture line going under the optical sensor to measure and collect the spectral signals that reflect spray drift droplets along the line. The system prototype was tested for detection of small (Very Fine and Fine), medium (Medium), and large (Coarse) droplets within the droplet classifications of the American Society of Agricultural and Biological Engineers. Laboratory testing results indicated that the system could detect the droplets of different sizes and determine the droplet positions on the droplet capture line with 100% accuracy at the wavelength of 610 nm selected from the 6 bands to detect the droplets.