UV-NDVI for real-time crop health monitoring in vertical farms
Wei Zhihao, Wei Fang
Abstract
This study aims to develop a plant health monitoring system suitable for vertical farms. It began by creating a multispectral LED with UVA and NIR light sources and an IoT device capable of controlling the LED spectrum. The IoT device integrates a camera with switchable filters and RGB CMOS sensor to simultaneously calculate SI-NDVI (single-image normalized difference vegetation index) and UV-NDVI (UV induce red chlorophyll fluorescence normalized difference vegetation index). A lettuce cultivation experiment was conducted on a planting shelf in a controlled environmental room, with two cultivation planters containing the same nutrient solution and red lettuce variety. SI-NDVI and UV-NDVI were calculated every three hours to compare differences. Results showed that both UV-NDVI and SI-NDVI trends were similar in monitoring lettuce growth. However, UV-NDVI was more sensitive to plant health than SI-NDVI. By inducing drought stress, the UV-NDVI indicator was able to detect water deficiency anomalies earlier than visual observation.