UAV Workflow Optimization for the Acquisition of High-Quality Photogrammetric Point Clouds in Forestry

Abstract

Three-dimensional modelling using photogrammetric point clouds derived from UAV-based aerial imagery is currently a popular topic in the scientific community. In particular, the use of image-based point clouds to enhance and update LiDAR DSMs is of growing interest in forest environments, i.e. as a future forest inventory method. Thanks to very high resolution imagery acquired via low-altitude UAV optical sensor payloads, very dense and accurate photogrammetric point clouds can be reconstructed through a triangulation process by means of photogrammetry software. In order to validate the use of image-based point clouds for their potential use in operational forestry, further comparison studies with LiDAR DSMs are being carried out by various research institutions. The acquisition of UAV-based aerial imagery, with the aim of producing accurate photogrammetric point clouds, though cost-effective, is not without its challenges. Due to constraints regarding power capacity and fair weather windows, we came to develop an effective image acquisition workflow with an emphasis on precision flight planning. The aim of this paper is to explore the process of UAV-based aerial imagery acquisition for the purpose of producing photogrammetric point clouds, as well as to give an overview of the initial stages of our research. With the aid of an image acquisition workflow that is adaptable to various field conditions, technical failures and precision flight planning, we estimate that the acquisition of aerial imagery for point cloud production will become more efficient as well as more precise, and in turn influence the accuracy of the 3D-modelling of forested areas.