Assessing the feasibility of low-cost unmanned aerial vehicle data collection and automatic tree detection to determine silviculture stand stocking status in BC

Abstract

The objective of this research was to evaluate the capabilities and limitations of low-cost UAV-collected silviculture data derived from Digital Aerial Photogrammetric point clouds. Four regenerating forest stands with a mix of planted spruce and lodgepole pine were surveyed using traditional ground plot surveys, aerial plot surveys, and aerial wall-to-wall surveys. A total of 22 systematic grid sampled ground plots were used as the baseline data to compare survey accuracy and efficiency of survey methods. Aerial survey imagery was processed using Pix4D software to produce orthomosaics and elevation data of each stand, which was used by a custom QGIS algorithm to identify stand stocking metrics. [...]