Growth Diagnosis and Yield Prediction in Lettuce Using a Drone Equipped with a Multispectral Camera

In open-field vegetable cultivation, soil fertility and abiotic stresses substantially affect growth. Therefore, growth diagnostic techniques considerably facilitate accurate yield estimation and stable shipping. Fertilizer application and post-rainfall moisture damage affect lettuce growth in open fields, resulting in frequent fluctuations in supply. Additionally, lettuce is often used as fresh food; however, its short shelf life in cold storage poses challenges for managing shipments. Thus, the aim of the present study was to clarify if growth diagnosis and yield prediction in lettuce can be achieved with a normalized difference vegetation index (NDVI) acquired by a drone equipped with a multispectral camera. In trials with four fertilization levels, the yield, growth parameters, and total nitrogen content differed based on the amount of fertilizer applied. Aerial images of the test plots were captured using a drone equipped with a multispectral camera in the early stage of heading, and NDVI values were calculated from the acquired images by extracting only the lettuce vegetation areas. In a regression analysis between NDVI and yield, the coefficient of determination (R2) was 0.87, while NDVI and total nitrogen was 0.77. The R2 values were similar to or higher than those reported previously and were considered practical. Subsequently, a 48-h flood treatment was applied at the base of the lettuce plants to simulate wet damage caused by heavy rainfall. When compared under standard fertilization conditions, the waterlogged treatments produced lower yield and growth parameters than the non-flooded treatments. The values were comparable to those of the non-flooded treatment under no fertilizer conditions. NDVI values were calculated over time from before waterlogging treatment to harvest. Following treatment, there was an immediate decrease in NDVI values that persisted until harvest. Therefore, the NDVI value may serve as a growth indicator in stress conditions. In conclusion, it is feasible to diagnose growth and estimate yield reflecting nitrogen nutritional status by calculating NDVI values obtained from aerial images captured using a drone equipped with a multispectral camera during the early stage of lettuce heading.