REAL-TIME DISEASE DETECTION AT THE EDGE: AN IOT FRAMEWORK FOR PRECISION AGRICULTURE WITH LEAF IMAGE ANALYSIS

Abstract

An IoT-based crop management system is proposed that tackles limitations of traditional cloud-centric approaches. This system integrates edge computing directly with leaf imaging sensors for enhanced disease detection and in site soil nutrient analysis. Edge devices deployed near sensors process data from various sources, including soil moisture sensors, temperature sensors, and potentially additional sensors measuring NPK values and ph. This initial analysis on the edge devices, alongside processing from the image sensors, reduces the amount of data sent to the cloud. This enables faster response times for critical situations like rapid disease spread and changes in soil nutrient levels. High-resolution leaf images are captured for more precise disease identification compared to traditional methods. Machine learning models on the cloud platform then analyse the captured images and processed sensor data to identify diseases, assess soil fertility, and recommend actions based on the combined insights. This approach offers earlier disease detection, improved resource management for water, fertilizer, and pesticides, and ultimately increased crop yields. By combining edge computing with image sensors and machine learning, the system contributes to a future of smarter and more sustainable agriculture with efficient and automated disease detection and soil analysis.