AI-enabled IoT-based pest prevention and controlling system using sound analytics in large agricultural field

Abstract

Modern agriculture is more scientific, analytical, and technological. IoT technology has smartened agriculture. This upheaval changed farming and created many opportunities. However, farmers contend with extreme concerns that lead to production losses. Pests are the most significant contributors to the economic losses that occur in agriculture. Because of this, farmers extensively use pesticides to combat pests that affect plant life. To address this issue, a study proposes the implementation of an AI-enabled system that uses real-time data from the IoT and advanced analytics technologies to detect, prevent, and control pests. The primary objective is to reduce the dependence on pesticides and minimize their harmful impact on our surroundings. The recommended method uses a PID infrared sensor to detect the pests using body heat. Sound analytics records pest sounds to confirm their existence in the field. Once sound analytics and a PID sensor confirm the existence of pests, an ultrasonic generator produces ultrasonic waves that repel pests from the agricultural field. The feature matrices and statistics measures were obtained from 1750 pest sounds using LPCC algorithm. The data was then trained, validated, and tested using DenseNet system models. The proposed system generated ultrasonic sound at multiple frequencies. These frequencies affect pests' hearing, making them uncomfortable in their surroundings. The proposed DenseNet model achieved pest detection accuracy of 99.65%, a sensitivity of 99.53%, a specificity of 98.06%, a recall of 98.89%, a precision of 99.29%, and an F1 score of 98.87%, which is superior to the ANN, SVM, RCNN, YOLO, and KNN techniques that are currently in use. The average design efficiency for preventing and controlling pests has achieved 98.87%.