Project Title: AgroScan AI: Autonomous Edge-Computing Rover for Precision Farming
Description:
This project addresses the critical problem of indiscriminate pesticide application and watering in modern agriculture, which leads to severe environmental pollution, soil degradation, high operational costs, and water waste. The proposed solution is based on developing "AgroScan Al," an autonomous, low-power autonomous vehicle powered by a Raspberry Pi 5 and integrated with computer vision technologies.The core of the system utilizes a custom-trained YOLOv8 object detection model, deployed entirely offline (Edge Computing), to instantly classify plants into either beneficial crops or harmful weeds. To ensure high precision and overcome hardware limitations, the robot is programmed using a Finite State Machine (FSM) architecture, incorporating a 'Stop-and-Stare' scanning mechanism and dynamic turning algorithms to counteract physical friction. A Time-of-Flight(ToF) distance sensor provides real-time collision avoidance and precise targeting. Upon detecting a weed, the rover autonomously halts, verifies the target, outputs the decision to an onboard OLED display, and simulates targeted spraying before executing an evasion maneuver to prevent redundant application.The expected outcomes include a substantial reduction in pesticide consumption, lowered agricultural operating costs, and the mitigation of chemical runoff intogroundwater. Furthermore, the project demonstrates the feasibility of deploying advanced Al models on low-cost, offline edge devices, providing an easily adaptable and sustainable precision farming method tailored for water-scarce and resource-limited environments.
Uploaded At: 2026-07-05 03:15:31