2012 Fiscal Year Final Research Report

Construction of airborne asbestos fibers automatic counting system incorporated the neural network technique

Research Project

Project/Area Number	22560585
Research Category	Grant-in-Aid for Scientific Research (C)
Allocation Type	Single-year Grants
Section	一般
Research Field	Architectural environment/equipment
Research Institution	Osaka University
Principal Investigator	INOUE Yoshio 大阪大学, 工学研究科, 助教 (60203262)
Co-Investigator(Kenkyū-buntansha)	KONDO Akira 大阪大, 工学研究科, 教授 (20215445)
Research Collaborator	SAKAMOTO Minori , 兵庫県環境研究センター, 主任研究員 (20446818) KOSAKA Hiroshi , 元兵庫県環境研究センター, 主任研究員 SATO Yukihiro , 元大阪府立産業技術総合研究センター, 主任技師
Project Period (FY)	2010 – 2012
Keywords	アスベスト / 繊維状物質 / 画像処理 / 自動計測 / 位相差顕微鏡 / 偏光 / ニューラルネットワーク
Research Abstract	An automated system to measure the fiber concentration of airborne asbestos fibers in the atmospherewas constructed in this study. The new technique is able to distinguish between asbestos fibers and non-asbestos fibers by using image analysis techniques and neural network concepts. It is difficult to distinguishing Amosite fibers and Glass fibers by the conventional method, which use geometric information of the fibers for the identification process. The efficiency and the accuracy of identifying asbestos fibers by the new method were improved by adding polarization information to the identification process.

(6 results)