Analysing dynamic processes in range data sequences

In this thesis a technique to estimate dynamic processes in range data sequences is developed. This includes the instantaneous velocity field (range flow) of a deformable surface and local expansion rates. For the velocity estimation novel differential constraint equations for the depth and intensity data are introduced. These constraint equations are then combined in a general total least squares parameter estimation framework. It turns out that this method can be used for a much broader class of problems where the parameters describing dynamic changes in multi-dimensional data are to be estimated. In addition to a confidence measure does the algorithm yield type measures indicating whether and to what degree there are linear dependencies in the data. Due to these dependencies the full parameter (range flow) set can usually not be computed at all observed data points. To overcome this a special regularisation scheme is developed that makes use of the determined local data structure. Surface expansion rates can then be computed locally from such regularised range flow fields. After an accuracy analysis of the presented algorithms they are applied to study living castor bean leaves. It is shown that this method can be used to investigate the movement and growth of such leaves with high spatial and temporal resolution.