Application of infrared thermography for irrigation scheduling of horticulture plants

Abstract

Infrared thermal imaging has acquired remarkable acceptance in the agricultural field for numerous applications ranging from irrigation scheduling, crop yield estimation, plant disease detection, fruit maturity evaluation, bruise detection to nursery monitoring. This paper presents a practical application of thermal imaging for irrigation scheduling of horticulture plants under different irrigation treatments. Four similar hibiscus potted plants were considered under different water conditions (slight, mild, moderate, and severe water stress). The thermal images were obtained from different canopy zones (front view, top view, left side view, right side view, and front angle view) using a ground-based hand-held thermal camera. The main objectives of the present study include (i) finding the effect of different irrigation treatments on potted plants and a corresponding change in their thermal signature, (ii) finding a relation between plant-based variables and soil-based variables, and to further develop a threshold value for irrigation scheduling (iii) to compare and simplify the calculation of the temperature of the canopy (Tc) and water stress indices namely crop water stress index (CWSI) and stomatal conductance index (Ig) using various reference and non-reference thresholding techniques and to improve their diagnostic accuracy. We can establish quantitative relationships between plant and soil-based parameters. The relationship between temperature of canopy (Tc) and temperature of soil (Ts) pair is best suited to find the plant under water stress. The threshold value pair (Tc and Ts) for the plants under study are 29.5°C (Tc) and 28°C (Ts) for plant A, 29.5°C (Tc) and 27.5°C (Ts) for plant B, 29.5°C (Tc) and 27°C (Ts) for plant C, and 29.5°C (Tc) and 26.5°C (Ts) for plant D. The most suitable method in terms of applicability and scalability, is histogram gradient thresholding (HG) method to determine Tc accurately. The critical m (for standard deviation envelope thresholding SDE method) and ratio of pixel change (RPC) (for HG method) value estimated can be considered an important indicator of water stress in the plants. Among different approaches to compute water stress indices, simplified and HG methods have minimum root mean square error (RMSE) for plants A, B, C, and D compared to statistical and SDE methods.