Ecoinformatics, Biometrics and Forest Growth

PhD topic D6


Design and implementation of an intelligent data interface between an electromagnetic 3-d digitizer and a graph-based modelling software


Description:


As an alternative to utilize laser scanning for the acquisition of the 3-d structure of vegetation, there is a device consisting of an electromagnetic sender and a receiver which permits to track the 3-d coordinates of every point in space which is marked using a sort of "3-d pen" (Polhemus Fastrak digitizer). This sort of device is used also in Virtual Reality. It shall be used to acquire points from plant shoot systems and leaves, and also from root systems which are digged out from the ground, in order to have data as a basis for simulation models of plant growth and structure to be used in agronomy and forestry. The intended information flow is from the digitizer to a dedicated graph-grammar based modelling platform, GroIMP (Kniemeyer 2008).


Using the digitizer requires some skills and some caution (e.g., no metal objects have to be present nearby). To guide the person who makes the measurements, an "intelligent" software would be useful which precomputes the structure which can already be reconstructed from the existing data, and which helps organizing the next steps for data capture (finishing the branch, inserting missing diameter values, marking positions of leaves, fruits etc.).


Together with the digitizer, Polhemus provides also a software, but this has a too poor functionality. A better software, PiafDigit, was developed at INRA (France), but its user interface is only in French and it has obviously memory leakages which makes it unstable. Functionality is still restricted. The new product should avoid these problems and should provide an extended set of features, including the possibility to capture supporting points for spline interpolation of leaf and flower shapes. Certain patterns should be detected automatically in the data. Shortcuts should be available for the user, e.g. for frequently repeated patterns, to speed up the acquisition process.


Furthermore, the requirements of this "intermediate software" are different from that of GroIMP (where the data finally should be stored), because it has to be used interactively during the process of measurement and thus has to be very quick and dedicated to its purpose. Nevertheless, an interface to GroIMP has to be implemented as well. Following the processing pipeline for the 3-d data, the tools for handling sets of geometrical primitives and point clouds in GroIMP shall be improved. Tests shall be documented to demonstrate practical usability of the new software, including a full measurement campaign at a whole plant.
Potential applications go beyond the field of plant modelling, for instance the device can be used to capture the exact shape of real objects for their use in virtual environments, e.g. in games.


Cooperation partner and additional external adviser: Dr. Peter Surový, University of Evora (Portugal).


Literature: