Ice crystal classification based on silhouettes
H. Lindqvist*,1, K. Muinonen1,2 and T. Nousiainen1
- 1 Department of Physics, University of Helsinki, P.O. box 48, FI-00014 Finland.
- 2 Finnish Geodetic Institute, P.O. Box 15, FI-02431 Masala, Finland.
An automatic classification system for ice crystals is being developed based on shape parameters derived from crystal silhouette perimeters. As a test set, we use a small sample of crystal silhouettes captured by a cloud particle imager in cirrus clouds. Preliminary results show that the chosen parameters are suitable for separating single and aggregate crystals, and further, dividing these into plates, plate aggregates, single rosettes, and rosette aggregates. Columns and bullets remain as one unseparated group. The results are encouraging for future studies on larger observational data sets, but also offer a method for testing ice-crystal shape models against images of real ice crystals.
- * Corresponding author