The determination of daily concentrations of atmospheric pollen is important in the medical and biological fields. Obtaining pollen concentrations is a complex and time-consuming task for specialized personnel. The automatic location of pollen grains is a handicap due to the high complexity of the images to be processed, with polymorphic and clumped pollen grains, dust, or debris. The purpose of this study is to analyze the feasibility of implementing a reliable pollen grain detection system based on a convolutional neural network architecture, which will be used later as a critical part of an automated pollen concentration estimation system. We used a training set of 251 videos to train our system. As the videos record the process of focusing the samples, this system makes use of the 3D information presented by several focal planes. Besides, a separate set of 135 videos (containing 1234 pollen grains of 11 pollen types) was used to evaluate detection performance. The results are promising in detection (98.54% of recall and 99.75% of precision) and location accuracy (0.89 IoU as the average value). These results suggest that this technique can provide a reliable basis for the development of an automated pollen counting system.
19 September 2000 in International Journal of Biometeorology
The study assesses differences in the results retrieved by Hirst samplers of the same model in order to determine their sampling variability. Five seven-day Burkard spore traps were compared. Four were located at ground level at the same site, separated by less than 2 m from each other. The fifth was placed on a terrace 16 m above ground, close to the other four samplers. An automatic meteorological station at ground level close to the spore traps was working simultaneously. Sampling was done for 51 days. The daily and hourly pollen concentration data were compared and the pollen distribution over the microscope slide surface was analysed. There were highly statistically significant strong positive correlations between spore traps themselves, with equivalent measurements of the daily and hourly patterns. Differences in hourly values between samplers were greater than differences in daily values, and the hourly peaks did not always match among samplers, with typically lags of two or more hours between traps. Analysis of the surface deposition of pollen on the slides showed pollen loss towards the edges and pollen appearing even outside the slit intake width. As hourly peaks indicated a degree of variability, this type of data requires to be interpreted with relative caution.