Evaluation of UAS-Based Remote Sensing for Measuring Forage Properties at an Experimental Grazing Land

This study uses Unmanned Aircraft Systems (UAS) to provide multispectral imagery and point cloud data to measure the health, height, and structure of plants. The high spatial resolution coupled with control over the temporal resolution of the data provide a potentially invaluable tool for grazing land management. Study areas were flown throughout the growing season in 2016 and 2017 at the Texas A&M AgriLife Beeville Station at the Tifton 85 and Sandy fields to measure forage health by calculating four vegetation indices (VIs) from the resulting reflectance images. VIs were used to estimate forage parameters such as plant height, herbage mass, and protein content, which were measured in the field for comparison. Digital elevation models and digital surface models were also created from the structure-from-motion (SfM) point cloud generated from the UAS flight. Multiple filtering algorithms were used to classify the raw point cloud into ground and non- ground classes. The classified points were then used to create ground and surface models for the 2016 flights. These surfaces were used to create terrain maps and to attempt to estimate plant height throughout the pasture. Linear regression analysis was performed to determine the strength of the association of each VI to forage field measurements for the 2016 and 2017 flights. Herbage mass and crude protein show the most significant regression with most of the vegetation indices, especially in the 2017 flights. Inconsistency between the 2016 and 2017 flight results highlights the importance of flight parameters in the resulting data. Future field data collection may show stronger relationships given improvements in experimental design and enable more comprehensive and accurate estimates of grazing land forage quantity and quality.