Remote determination of disease, phenology, phenotype, and nitrogen on the wild blueberry field

Abstract

Utilizing remote sensing for research and development is essential in enhancing site-specific management practices and estimations of wild blueberry field characteristics. This research aimed to address challenges with site-specific management practices by enhancing productivity, promoting sustainability, reducing production costs, and minimizing environmental impact through decreased agrochemical use. This was achieved partly by identifying plant phenotypes, phenology, and early detection of Monilinia and Botrytis floral diseases, and nitrogen use. An increase in N significantly improved plant growth due to the perennial nature and potential nutrient carryover in wild blueberries. Effective estimations of LNC and LAI were achieved using VIs. Further monitoring and estimation of the growth and development parameters of the plant revealed that LAI, floral, and vegetative bud stages can be estimated at the tight cluster (F4/F5) and bloom (F6/F7) stages with R2/Lin’s CCC values of 0.90/0.84, respectively, although there were challenges in estimating floral and vegetative bud numbers. Additionally, NDVI, ENDVI, GLI, VARI, and GRVI significantly contributed to achieving the predicted values, while NDRE had minimal effects. A pixel classification method successfully identified Vaccinium angustifolium f. nigrum, a disease-susceptible phenotype, with an overall accuracy (OA) of 80%. Estimating the incidence and severity of Monilinia and Botrytis blight on the field posed a challenge, although, the VIS-VIs performed better compared to the NIR-VIs. Classification assessment using hyperspectral data showed that discrimination of MB and BB disease from healthy plants was achieved with an OA of about 96.6% using an SVM or RF classifier. This influences production costs by adopting a spot application of fungicides rather than a blanket application. These findings underscore the utility of remote sensing in discerning floral diseases, assessing phenology, identifying phenotypes, and monitoring nitrogen utilization in wild blueberries.