Robotized Early Plant Health Monitoring System

Abstract

Imagining analysis tools, and gas sensors are becoming more frequently integrated into smart systems for early plant disease detection. Multispectral and hyperspectral imaging techniques are commonly used for disease detection in plants and vegatitve produce. However, there is still a limitation in the capability of designing systems for the detection of diseases in real time under field conditions. This paper introduces the development of an automated system that has the capabilities of monitoring plant health and stresses under field conditions in real time. A robot is designed and implemented to provide field based analysis of plant health using visible and near infrared spectroscopy. Fusion of visible and near infrared images are used to calculate the Normalized Deference Vegetative Index (NDVI) to measure and monitor plant health. The robot is designed to have the functionality of moving across a specified path within an agriculture field and provide health information of leaves as well as position data. The system was tested in a tomato greenhouse under real field conditions. The developed system proved effective in accurately classifying plant health into one of 3 classes; underdeveloped, unhealthy, and healthy with an accuracy of 83%. A map with plant health and locations is produced for farmers and agriculturists to monitor the plant health across different areas. This system has the capability of providing early vital health analysis of plants for immediate action and possible selective pesticide spraying.