The spectral dimension of a point cloud -- finding crop marks in radiometrically calibrated ALS data

Abstract

Airborne laser scanning (ALS, also referred to as airborne LiDAR) is a widely used data acquisition method for topographic modelling. Due to its ability to accurately and densely sample the terrain surface it became a common used technique for the generation of digital terrain models. Compared to other DTM acquisition techniques, it especially excels in forested areas due its active direct 3D sensing principle. Small footprint ALS systems can penetrate the vegetation layer through gaps in the canopy and might therefore still receive a laser echo from the terrain surface even in densely vegetated areas. In archaeology, this potential has revolutionised prospection of forested areas. For the analysis and classification of the topography, geometric criteria derived from the acquired 3D point cloud are typically used. However, ALS systems deliver in addition to the 3D position the amplitude of each echo (often referred to as intensity). In contrast to standard discrete echo ALS systems, advanced full-waveform ALS systems allow digitising the whole return signal and hence enable to estimate the echo width of each acquired 3D point. Based on these additional physical observables the return power of the target can be calculated. In order to accurately study the radiometry acquired by ALS sensors a correction of point-wise influencing factors (e.g. range, angle of incidence, surface characteristics, atmosphere, etc.) based on in-situ reference surfaces has to be performed. This process of radiometric calibration enables to convert the amplitude and echo width into absolute radiometric values which describe the characteristics of the observed surface for the used laser wavelength. This contribution describes the whole workflow for the absolute radiometric calibration of ALS data. As a result, a surface dependent reflectance value per individual ALS echo resp. point is determined. Based on these values a single channel (ALS wavelength) true orthophoto (each radiometric value is on its 3D position) can be determined by the application of a certain interpolation method. Furthermore, it is important to mention that the reflectance values displayed in the orthophoto are independent from the ambient light (due to the active laser illumination) and therefore neither affected by sun shadows. Within this paper, the process of radiometric calibration and true orthophoto generation is demonstrated with a full-waveform ALS data set of the archaeological study site Carnuntum, Austria. The resulting true orthophoto will be presented and discussed from a technical and archaeological point of view. Furthermore, detected vegetation marks in the ALS orthophoto will be compared to vegetation marks that have been identified in simultaneously acquired passive optical imagery. The final section of this paper provides a short summary of the presented work and an outlook into future research work.

Michael Doneus
Michael Doneus
Key Researcher
Geert Julien Joanna Verhoeven
Geert Julien Joanna Verhoeven
Vice Director, Senior Researcher

Related