DEVELOPMENT OF AN AUTOMATIC VARIABLE RATE SPRAYING SYSTEM BASED ON CANOPY CHARACTERIZATION USING ARTIFICIAL INTELLIGENCE

Abstract

Spraying on tree crops must consider the canopy's structural features to maximize its effectiveness. The main drawbacks to VRI technology include the complexity of successfully implementing it and the lack of evidence that it assures better performance in net profit or water savings. Hence, a novel framework based on canopy characterization was presented in this research for an automatic variable-rate spraying system. The first phase was collecting the data, and the next was cleaning it to eliminate redundant information. The pre-treated data are then entered into the Crest- Stride-wise Regression Framework we devised, where we extract the canopy features and evaluate additional parameters. In addition, our proposed model automatically predicts the nozzle's flow rate and pressure based on a threshold value. Thus, this research shows that the recommended strategy achieves 99.98% accuracy, 99.99% precision, 99.99% F1 score, and 99.99% recall. As a result, our study enables safer and more efficient spraying distribution in the agricultural sector.