Translating bumblebee dialect…
This dataset represents annual maps of high-resolution vegetation canopy height of the Gothic Townsite area surrounding Rocky Mountain Biological Laboratory. Data is derived from photogrammetry processing of imagery collected by an Uncrewed Aircraft System (UAS, e.g. "drone"). Image data was captured by 40 megapixel RGB camera attached to a quadcopter aircraft flying at 100m above the ground surface. Image data has a nominal ground sample distance (pixel resolution) of approximately 1.25 cm, but the height products, produced using structure-from-motion photogrammetry have a resolution of 5 cm. Values represent the maximum height above the ground surface (in meters) of all objects that are attached to the surface. Per-pixel values represent maxima from 3-4 seasonal composites per year. Data is derived from 3-4 UAS flights per year occuring from late June to late August, representing conditions of low wind and even illumination. Flights were conducted approximately weekly within 3 hours of solar noon, in full-sun conditions where possible. The flights typically take approximately 75 minutes to complete, and flightlines were oriented NNW-SSE at a 28m spacing, starting in the western proportion of the imaging area and progressing east. This means that western parts of the imaging area were typically imaged approximately 1-hour beore the eastern edges. Raw images were captured as JPEGs at a rate of 1.4 images per second, providing approximately 80 percent forward overlap and 75 percent side overlap between adjacent images. For densely forested parts of the landscape, the forward speed of the aircraft was reduced to 8 m/sec to increase the forward overlap to 85%, improving orthomosaic performance for complex vegetation. Raw image data was georeferenced and orthorectified in Agisoft Metashape software (version 1.7), using information from the UAS system's Real-time Kinematic (RTK) global satellite navigation system as well as 8-15 ground-control points geolocated to within 2 cm. Independent assessment indicates a geographic precision of approximately 5 cm (Mean Absolute Error). Photogrammetry reconstruction generated dense point clouds with approximately 4.1 billion points per flight using Metashape's high quality settings. A digital surface model was generated for each flight using the Metashape default algorithm with interpolation disabled. DSMs from individual flights in a given year were composited to generate a map, taking the pixel value of maximum height across DSMs. Remaining areas without height values (<1% of the imagery area) were replaced with a digital elevation model value. Ground surface (DEM) height values were subtracted from the DSM to yield a canopy height map. The DEM was derived from October 2021 UAS imagery processed using the above workflow. Ground points were classified using a sequential cloth simulation filter and greenness filters in CloudCompare software, and additional manual curation work removed additional non-ground points. The 25 cm resolution canopy height maps are suitable for per-pixell time-series analysis due high accuracy (5 cm) in georeferencing and DSM reconstruction. An alternative version of this dataset, delivered at 5 cm resolution, has higher resolution but less stable per-pixel trends.
Items connected by shared entities, co-authorship, citations, or semantic similarity.