2018 ASHS Annual Conference

With rapid advances in unmanned aerial systems (UAS) and automated multispectral sensing, agriculture professionals are adopting these technologies to manage their operations. Multispectral imaging data in the wavelengths of 550 nm, 650 nm, 709 nm and 820 nm were collected from three commercial fields of roses in California. The automated UAS classified the images based on normalized vegetation difference indices (NDVIs) and identified regions of the fields that appeared stressed. Field verification of these areas indicated that there were interactions between abiotic and biotic stress types. The three main areas identified using multispectral imagery and field verification included: (1) a rapidly drained area with coarse textured soils and rose plants affected with rose rosette virus (RRV), or powdery mildew (Sphaerotheca pannosa var. rosae), (2) an area with excess moisture and drainage issues with a heavy cultivar specific rust infection and two other varieties affected with rose mosaic disease and, (3) a cultivar of Rosa rugosa affected with grey mold (Botrytis cinerea) and minor infections of Cercospora leaf spot (Cercospora rosicola). Here, we show that an automated UAS can identify regions of a field where a combination of poor site factors and disease exist. Future work will involve evaluating multispectral imagery in differentiating between individual biotic stress types and monitoring rose plant health and cultivar-specific flower production.