科研之友
微信
新浪微博
Facebook
分享链接
摘要
Rapid detection of airborne fungal and bacterial spores would enable public agencies to respond quickly and appropriately to intentional releases of hazardous aerosols. Automated analysis of microscope images and automated detection of near-monodisperse peaks in aerosol size distribution data offer complementary approaches to traditional methods for the identification and counting of fungal and bacterial spores. First, spores of the fungus Scopulariopsis brevicaulis were aerosolized in a chamber and then collected with a slit impactor; later, digital microscope images were analyzed manually to determine spore cluster distributions. The images also were analyzed with ImageJ, a program that automatically outlined objects and measured Feret's diameter, area, perimeter, and circularity. These characteristics were used to identify spore clusters automatically using two data analysis methods. Second, a computer program was developed to discriminate near-monodisperse bioaerosol peaks from those for polydisperse ambient particulate matter (PM) and was successfully tested using simulated and real aerosol mixtures. The observed agreement between manual and automated spore counts and the ability to detect spore peaks suggest that it may be possible to develop a system to recognize intentional releases rapidly through examination of particle morphology and size distributions. The peak detection procedure is potentially the fastest technique when used with real-time instrument data, but assumes that intentional releases would consist of large numbers of uniformly sized particles in the respirable size range.
- 出版日期2012
