SAR radargrammetry and scanning LiDAR in predicting forest canopy height

Our objective was to evaluate the accuracy of estimating forest canopy height when using scanning LiDAR and TerraSAR-X stereo radargrammetry. The study area was located in southern Finland. We used SAR radargrammetry and LiDAR to extract 3D point clouds to derive predictors used in the non-parametric prediction of forest canopy height. We used tree-wise measured field plots (n=110) as reference data. Our results showed that with SAR radargrammetry, the relative RMSE for forest canopy height was 12.2% whereas it was 8.1% with LiDAR. We concluded that SAR radargrammetry is a promising remote-sensing method for predicting forest canopy height when an accurate digital terrain model is available.