• Project number: F 2479
• Institution: Federal Institute for Occupational Safety and Health (BAuA)
• Status: Completed Project

Description:

The inhalation of fine and ultra-fine dust particles in the living and working environment can have a negative impact on human health. Therefore, it is necessary to reliably determine the composition of dusts and the concentration of critical components. There is currently little reliable information on the composition of particles indoors, but this is most relevant for overall human exposure due to the long duration of stay. In addition, the necessary investigations for the characterisation of aerosols in indoor spaces are very time-consuming and personnel-intensive, making evaluation automation urgently necessary.

For this reason, the research project funded by the German Environment Agency (Umweltbundesamt, UBA) developed an automated evaluation system based on a combination of several imaging and chemical-analytical microscopy methods. This provides information on both the shape and the chemical composition of the fine dust particles, which can then be identified using computer-assisted methods.

The starting point for the analysis is formed by electron microscope images of filter surfaces used to collect particles from the air at workplaces and other indoor spaces. In order to be able to assign them to known particle groups, the shape of the particles localised in the image is measured and their elemental composition and chemical properties are characterised using spectroscopic methods.

For this segmentation step, machine learning methods are used that were trained on dusts of known composition. The developed procedures were implemented in software. Both the automatic image generation and the automatic image analysis and classification enable the evaluation of a large number of particle images and spectra, which are a prerequisite for meaningful statistical studies. Thus, the successfully completed research project provides an important basis for an efficient assessment of indoor particulate pollution.