LIDAR strip adjustment: Application to volcanic areas

Abstract

DEMs derived from LIDAR data are nowadays largely used for quantitative analyses and modelling in geology and geomorphology. High-quality DEMs are required for the accurate morphometric and volumetric measurement of land features. We propose a rigorous automatic algorithm for correcting systematic errors in LIDAR data in order to assess sub-metric variations in surface morphology over wide areas, such as those associated with landslide, slump, and volcanic deposits. Our procedure does not require a priori knowledge of the surface, such as the presence of known ground control points. Systematic errors are detected on the basis of distortions in the areas of overlap among different strips. Discrepancies between overlapping strips are assessed at a number of chosen computational tie points. At each tie point a local surface is constructed for each strip containing the point. Displacements between different strips are then calculated at each tie point, and minimization of these discrepancies allows the identification of major systematic errors. These errors are identified as a function of the variables that describe the data acquisition system. Significant errors mainly caused by a non-constant misestimation of the roll angle are highlighted and corrected. Comparison of DEMs constructed using first uncorrected and then corrected LIDAR data from different Mt. Etna surveys shows a meaningful improvement in quality: most of the systematic errors are removed and the accuracy of morphometric and volumetric measurements of volcanic features increases. These corrections are particularly important for the following studies of Mt. Etna: calculation of lava flow volume; calculation of erosion and deposition volume of pyroclastic cones; mapping of areas newly covered by volcanic ash; and morphological evolution of a portion of an active lava field over a short time span.