On intensity-based stochastic models for terrestrial laser scanners
Geodesists pride themselves to have sound knowledge of the stochastic characteristics of various established sensors, ranging from total stations to levels. This information is for instance used to thoroughly control the influence of individual observations within an adjustment, to reveal outliers or to identify statistically significant deformations. The emergence of terrestrial laser scanners (TLS) have clearly initiated a paradigm shift regarding new possibilities of data acquisition and consequently allowed to conquer novel fields of application, such as forest inventory or Palaeontology. Yet, the aforementioned self-conception clearly had to be questioned, as for more than a decade no meaningful stochastic models have been provided for these upcoming sensors – neither by the manufacturers of terrestrial laser scanners nor the scientific community. This circumstance is rather astonishing since appropriate stochastic models are of vital importance for tasks such as sensor calibration of terrestrial laser scanners, registration of point clouds or direct georeferencing of TLS. Hence, this talk will show why the established approaches in stochastic modelling cannot work and how it can be achieved using the radiometric information captured by the scanner.