AI 'Agribot' Deployed to Spot Locust Swarms with Near-Perfect Accuracy

Destructive locust swarms pose a significant threat to agricultural production worldwide. To combat this, scientists have engineered an autonomous 'Agribot' that patrols fields, acting as a vigilant scout. This system integrates the Internet of Things (IoT) for automation—connecting sensors and a central Android application—with powerful deep learning models for visual detection.

At its core, the Agribot uses pre-trained Convolutional Neural Networks (CNNs), a form of AI that processes visual data, to analyse live video streams from the field. During testing, one specific configuration, using the VGG19 model with other machine learning methods, achieved a remarkable locust detection accuracy of 99.51%.

While moving at a steady pace of 0.3 metres per second, the robot demonstrated strong feasibility for real-time use. The system's user-friendly design also earned a high usability score of 86%. Although researchers noted some limitations during implementation, the Agribot represents a promising technological leap in the proactive defence of our food crops.