Forest variable estimation using a high-resolution digital surface model

Abstract

Airborne laser scanning is today considered the most accurate remote sensing method for forest inventory. The main advantage of laser scanning is the three-dimensional data. Three-dimensional canopy surface models can also be derived by means of digital aerial photogrammetry on the basis of optical remote sensing imagery. The photogrammetric surface models require high-resolution aerial images with stereo coverage. In this study, both a canopy height model derived from a photogrammetric digital surface model and laser point data were tested in estimation of sample-plot-level forest attributes. The attributes tested include diameter, mean and dominant height, basal area, and volume of growing stock. The results indicate that the laser data give higher accuracy for the estimated forest variables than does the photogrammetric canopy height model. The stand dominant height was the most accurately estimated variable from both data sources and showed the smallest difference between the laser data and photogrammetric canopy height models. The performance of the photogrammetric model was poorest in estimation of basal area and volume of growing stock.