Object-based spatiotemporal analysis of vine canopy vigor using an inexpensive unmanned aerial vehicle remote sensing system

Remotely sensed imagery provides a rapid assessment of spatial variability in grapevine canopy vigor that correlates with crop performance. Unmanned aerial vehicles (UAVs) provide a low-cost image acquisition platform with high spatial and temporal resolutions. Using a UAV and digital cameras, aerial images of a Texas vineyard were captured at postflowering, veraison, and harvest. Imagery was processed to generate orthophotos in units of reflectance, which were then segmented to extract per-vine estimates of canopy area (planimetric extent) and normalized difference vegetation index (NDVI)-based canopy density. Derived canopy area and density values were compared to the harvest variables of number of clusters, cluster size, and yield to explore correlations. Planimetrically derived canopy area yielded significant, positive relationships, whereas NDVI-based canopy density exhibited no significant relationships due to sensor-related radiometric inaccuracy. A vine performance index was calculated to map spatial variation in canopy vigor for the entire growing season. Future management zones were delineated using spatial grouping analysis.