Analysis of Crop Dynamics through Close-Range UAS Photogrammetry

Food security, in the context of a growing world population and spatial restrictions for farmlands, demands new approaches to optimize crop productivity. This work proposes a methodology for monitoring agricultural fields over time with high spatial and temporal resolution, using low-cost RGB sensors onboard of small scale drones in combination with computer vision and programming techniques. The methodology was applied to study bean crop dynamics in Costa Rica over a six-week period with weekly observations. Resolutions up to 8.5 mm/pixel and RMSE values in the millimeter range could be achieved from dense point clouds and digital elevation models, which allow the observation of crop progression and decay with an accuracies in the cm-range.