Knowledge acquisition was done by REA ( UCO and UAB) in collaboration with INRIA (Orion).
Only the experts in palynology are able to recognise the class (family, genus, specie, group) of a given pollen grain. For that, they use their general knowledge in palynology such as pollen taxonomy, schematic representations of the various pollen classes in terms of their sub-parts or their characteristics, and the pollinic calendars depending on the geographical location. An automatic pollen recognition system has to be guided by this knowledge.
The pollen knowledge used to identify the pollen grains come from:
The system tries to mimic the palynologists. For this, pollen knowledge is used to identify each grain.
We have developed an original interface, named annotate, that allows the expert to look into a 3D image sequence and annotate it with texts and drawings to indicate key characteristics for the recognition.
The flowering period is an important information for the pollen recognition process. Knowing the sampling date for a pollen grain can help to identify its type.