Correction pit free canopy height model derived from LiDAR data for the broad leaf tropical forest

Abstract

The accuracy of forest inventory information, such as tree height estimation and individual tree crown delineation depending on the extraction of Canopy Height Model (CHM) that represent the absolute tree height of vegetation. However, uncertainties variation of the CHM known as a pit, negatively influence the biophysical measurements. These occur when the laser beam penetrates the part of the leaf and the branch before produces the first return to represent the surface of canopies. Despite of that, multiple laser beam from the different flight line cause a variation of the tree heights. To reduce the error of CHM, this work applied pit free algorithm CHM for broad leaf tropical forests. The results shown that, the pit free CHM produces a good correlation with the ground data with R2 0.86, meanwhile, the standard CHM with R2 0.78. For the accuracy assessment, Accuracy Index (AI) has been used to calculate the error of commission and omission for tree detection. The analysis proved that the pits free CHM improved the accuracy of tree detection with the total Accuracy Index value 69% and standard CHM recorded 52.3%.