Forest canopy assessment through growing seasons; a low cost approach using hemispherical photography, unmanned aerial vehicles (UAVs) and Structure-from-Motion (SfM)

Abstract

Assessing biophysical forest parameters such as Leaf Area Index (LAI) and canopy height is of vital importance for monitoring forest productivity. This study provides an evaluation of canopy parameter retrieval through the forest growing season using a combination of two low cost digital optical imagery techniques. Fieldwork was conducted in a small forest stand in Dryden Farm near Edinburgh, at three development stages from April to June 2016. Hemispherical photography from below the canopy was used to calculate LAI and light transmission. These were compared to canopy area estimated using UAV imagery from the top, suggesting that UAVs can be used to assess changes in leaf area through time. The Structure-from-Motion (SfM) technique from the UAV images was evaluated for creating 3D models. Their accuracy was calculated using field measurements as a validation method. The results of this study showed that SfM performed well for producing accurate Digital Surface Models (DSMs) only during leaf-on acquisitions (resolution 4.86cm pixel-1) due to the increased penetration in conditions without foliage. The DSM was used to assess the SfM method for estimating tree heights. A significant strong positive correlation was observed between the model derived canopy heights and field measurements (R2=0.96, RMSE=81cm). The Ground Control Point (GCP) method assessed the horizontal accuracy and produced a RMSE of 1.37m, which is considered suitable for forestry applications. This study concludes that a combination of these imagery techniques can be utilised to estimate and compare forest stand parameters through and between growing seasons with adequate accuracies, at very low costs.