An algorithm and a device for counting airborne pollen automatically using laser optics

Airborne pollen is important in relation to the social issues of pollinosis and of the environmental effects of genetically modified plants. Existing methods for pollen counting involve counting and classifying the grains that adhere to a sampling surface, requiring much time and skilled labor. We therefore have developed a method of automatically monitoring pollen, using a laser-optics instrument. In this instrument, the sideways and forward scattering of laser light by each particle is recorded in real time for computer processing. A field experiment was conducted in 2005, comparing our method with that of the older Hirst method. A scatter plot was made of the forward scattering vs. the sideways scattering for each particle. An algorithm was developed to find the optimum rectangular region of the plot for each type of pollen, and a count of points inside this region was taken as the count for that type of pollen. For the three most common types of pollen found in the field test (Urticaceae, Poaceae, and Ambrosia), the daily counts from this algorithm were compared with the daily counts from the Hirst-type (Burkard) sampler. There was a very high correlation (determination coefficient approximately 0.8) between the results of the two methods.