3D pollen image recognition based on palynological knowledge

Authors
A. Boucher, P. Hidalgo, M. Thonnat, J. Belmonte and C. Galan

Abstract:

A pollen recognition system is developed using 3D pollen images and palynological knowledge. These two aspects represent the main originality of this work with respect to other existing systems. This system tries to imitate the methodology applied by the palynologist during the pollen count and identification. For each pollen grain, a sequence of 100 color images is taken from a light microscope with 60x objective, showing the pollen grain at different focus (with a step of 0.5 microns). To accomplish the recognition, it is necessary to work on 3D images and to use deep palynological knowledge. This knowledge describes the pollen types according to their main visible characteristerics and to those which are important for recognition. The strategy used for recognition first detects the pollen grain boundaries in 3D. Then, the main characteristics are searched for according to their location. For example, the top surface is analyzed to understand the ornamentation while the central images help to identify the cytoplasm or the exine thickness. The detection algorithms work on 2D images and validate their result in 3D (i.e. a pore is detected and validated if it appears in several images). Such algorithms use image processing methods like thresholding, gradient analysis, region segmentation, color analysis. Different algorithms are developed to identify a single characteristic with different appearances (a pore is seen differently from the polar view than the equatorial view). At present, some structures of the pollen grains can be understood and identified like the pore with annulus in Poaceae, the reticulum in Olea and similar pollen types or the cytoplasm in Cupressaceae. Currently, the system is developed using a database of 350 pollen grains (35000 images) with 10 to 20 sequences per pollen types. The preliminary results show the recognition of some pollen types, like Urticaceae or Poaceae or some groups of pollen types, like reticulate group.

Keywords:

BibTeX reference: