Automatic Counting Robot Development Supporting Qualitative Asbestos Analysis – Asbestos, Air Bubbles, and Particles Classification Using Machine Learning –

Abstract

Asbestos, particle, and air bubble counting generally supports qualitative asbestos analysis, using such procedures as dispersion staining. Operators conventionally check and count asbestos fibers visually using a microscope – a difficult, time-consuming process. The microscopic observation robot we are automating to support qualitative asbestos analysis images fibers and saves them automatically to a database. In this paper, we introduce image processing method using machine learning to count asbestos, particles, and air bubbles automatically.