M. Peter

Dense matching quality evaluation - an empirical study

The acquisition and processing techniques of Airborne Laser Scanning (ALS) data have improved rapidly in recent years. Due to the relative high costs of laser scanning, we want to explore the potential of detecting changes and updating Digital Surface Models using point clouds derived from Dense Image Matching (DIM). The prerequisite of this work is to evaluate dense matching quality. In this paper, a workflow is designed to evaluate dense matching quality using planar roof segments. ALS data are taken as reference. The workflow can be divided into two steps: roof detection and quality evaluation. Two types of accuracy plots are depicted based on single DIM points and single segments. The experimental results show the point-to-plane residuals of DIM points are around 3 cm.