Accuracy assessment of LiDAR elevation data using survey marks

Abstract

Airborne LiDAR has become the preferred technology for digital elevation data acquisition in a wide range of applications. The vertical accuracy with respect to a specified vertical datum is the principal criterion in specifying the quality of LiDAR elevation data. The quantitative assessment of LiDAR elevation data is usually conducted by comparing high-accuracy checkpoints with elevations estimated from the LiDAR ground data. However, the collection of a sufficient number of checkpoints by field surveying is a time-consuming task. This study used survey marks to assess the vertical accuracy of LiDAR data for different land covers in a rural area and explored the performance of different methods for deriving elevations from LiDAR data corresponding to the locations of heckpoints. Normality tests using both frequency histograms and quantile-quantile plots were performed for vertical differences between the LiDAR data and the checkpoints, so the appropriate measures (the formula 1.96×RMSE or the 95th percentile) can be used for the vertical accuracy ssessment of the LiDAR data for different land covers. The results demonstrated the suitability of using survey marks as checkpoints for the assessment of the vertical accuracy of LiDAR data.