Is a Picture Worth a Thousand Measurements? A Comparison of Drone Systems and Data Processing Methods for Rangeland Vegetation Monitoring

Sustainable use of over 160 million hectares of rangelands in the central and western United States requires effective management of woody species (brush). Vegetation monitoring is a critical component of rangeland management and is necessary for determining when brush treatment (one of the most cost-shared and implemented conservation practices on grazing lands) is needed. Ground vegetation measurement of transects or plots is labor intensive; however small unmanned aerial systems (sUAS), or drones, may offer a cost-effective alternative. These systems acquire extremely high-resolution imagery that can be processed into very detailed orthophotos or 3-dimensional models using Structure from Motion (SfM) technology.

In this study, a shrub-dominated site on the USDA-Agricultural Research Service-operated Walnut Gulch Experimental Watershed in southeastern Arizona, USA, was imaged to assess the feasibility of using sUAS to supplement or replace traditional ground measurement of woody vegetation species cover. Two different drone systems (DJI Phantom 4 Pro and Phantom 4 RTK), equipped with conventional or real-time kinematic (RTK) GPS, were flown with RGB cameras above a permanent transect area (approx. 1.6 ha) over two consecutive years. Contemporaneous ground measurements were collected in the second year. Multiple sUAS data processing methods were compared to assess length of processing time, ease of use, and quality of output. Potential usefulness of resulting orthophotography and SfM point cloud data for species identification was also evaluated.