Development of Deep Learning-Based Variable Rate Agrochemical Spraying System for Targeted Weeds Control in Strawberry Crop

Agrochemical application is an important tool in the agricultural industry for protection of crops. with conventional sprayers results waste applied agrochemicals, which not only increases financial losses but also contaminates environment. Targeted agrochemical using smart control systems can substantially decrease chemical input, weed cost, and destructive environmental contamination. A variable rate spraying system was developed deep learning methods development new models to classify weeds accurately spray on desired target. Laboratory field experiments were conducted assess sprayer performance classification precise target three CNNs (Convolutional Neural Networks) models. The DCNNs (AlexNet, VGG-16, GoogleNet) trained a dataset containing total 12,443 images captured from strawberry (4200 spotted spurge, 4265 Shepherd’s purse, 4178 plants). VGG-16 model attained higher values precision, recall F1-score as compared AlexNet GoogleNet. Additionally recorded percentage completely sprayed (CS = 93%) values. Overall all experiments, performed better than GoogleNet real-time precision spraying. revealed that Sprayer decreased increase traveling speed above 3 km/h. Experimental recommended achieve high makes it more ideal application. It concluded advanced has potential spot agrochemicals field. reduce crop input costs pollution risks.