next up previous
Next: System architecture Up: A Semi-Automatic System for Previous: A Semi-Automatic System for

Subsections


Context of the project

The European Aeroallergen Network

Aerobiology had undergone a dramatic expansion in the last 10 years with the formation of many national pollen sampling networks. According to recent epidemiological studies, between 15-20 affected by pollinosis. Evidence suggest that this percentage is increasing because of changing environmental conditions such as pollution. The term aeroallergen has been created to resume in one word the particles dispersed through the atmosphere which are able to create allergy responses in sensitive people.

Nineteen European countries are integrated in the European Aeroallergen Network (EAN)/European Pollen Information (EPI). Altogether 188 monitoring units distributed throughout Europe are sending weekly pollen information to the national networks and, at the same time, these national networks are sending the pollen information to the European Coordinator Centre. Aerobiology currently operates having common standards for the collection, the analysis, the representation and the interpretation of the pollen counts data.

Pollen prediction

One of the most important goals of aerobiologists is to be able to forecast the start, duration and severity of the pollen season.

When several yearly data series are achieved, methodological approaches are employed to obtain an empirical mathematical model for each selected pollen type or vegetal species. To construct these models phenological, aerobiological (pollen counts) and meteorological parameters (mainly rain and temperatures from the previous months weather conditions) are considered and they have an statistics base.

As atmospheric pollen databases have not yet enough long series, up to the moment experience and skill of the aerobiologists play an important role in predictions.

Pollen recognition

In Europe the instrument most widely used to sample allergenic airborne pollen is the Hirst sampler. This is a suction sampler based on the basic impact principle. The capture of the pollen grains occurs by impact on a sticky surface that is then adequately prepared to be examined under the optical microscopy. The pollen grain is drawn into the sampler by suction from a vacuum pump. A vane tail keeps the sampler orifice facing the wind. The rate of suction is 10 litres/minute, which is estimated to be the normal respiratory flux of a person.

One slide per operating day is obtained from this sensor. The task of the palynologist technician is to "count" this slide, this means to recognise on it the pollen particles, to give every pollen a name (family, genus, specie, group) and to produce a slide pollen spectrum at the end for the given day. The results can also be expressed per hour, this means hourly aerobiological data. This level of aerobiological analysis is also very interesting, because plants shed the pollen at different periods along the day and this knowledge can help allergy sufferers.

The technician do not read the whole slide, which could involve a lot of working time, and so the slide reading is also a sample. As a general rule, accepted by most of the European Aerobiological Groups, almost the 10 the slide must be read. Even using this type of sub-sampling, pollen count involves a great deal of working time because operators take about three hours to do readings for each daily slide to obtain a measure of mean daily pollen concentration. If the pollen analysis is made to obtain the results expressed per hour, the working time used to do the slide readings is notably increased.

Not only because of the time required to obtain the pollen measurements from the sensor's samples but also because possible human errors of counting and identifying the pollen grains can occur, it is of enormous interest to develop a system capable to recognise the pollen grains and to count them per types, this means to make possible an automatic evaluation of the atmospheric pollen concentration.


next up previous
Next: System architecture Up: A Semi-Automatic System for Previous: A Semi-Automatic System for
Alain Boucher
2001-03-06