ニューラルネットワークを利用したデジタルカメラによる粉じん濃度測定の実用化  [in Japanese] THE PRACTICAL DENSITY MEASUREMENT OF SUSPENDED SPRAYED CONCRETE DUST BY USING ARTIFICIAL NEURAL NETWORK  [in Japanese]

Access this Article

Author(s)

Abstract

トンネル建設など坑内作業中に発生する粉じんは,建設作業員がそれを吸引することでじん肺症を引き起こす原因のひとつと言われている.平成20年より,半月に1回の粉じん濃度測定がトンネル施工管理者に義務化された.従来,粉じん濃度測定にはデジタル粉じん計がよく使用されるが,測定に時間がかかることや測定機器固有の換算係数を設定する必要があることからより迅速で簡単な計測法が望まれていた.そこで,著者らはデジタルカメラを用いて坑内に浮遊する粉じん濃度の測定方法を考案した.しかし,実用化に向けての課題も明らかとなってきた.本研究では,フラッシュ撮影画像から粉じんの反射画像のみを抽出したうえで統計分析し,その結果をニューラルネットワークに学習させることにより,従来法以上の精度を有する粉じん濃度測定方法を提案する.

  The suspended dust generated in the tunnel construction work is one of the causes for tunnel workers becomes Pneumoconiosis. Therefore, the ministry of Health, Labor and Welfare of JAPAN dictated the employer has a management duty to measure the density of suspended dust twice a month from 2008. Ordinary, the density measurement of suspended dust had done by the digital dust meter. However, the digital dust meter is expensive and it need the measuring time for 10 minutes. In addition, it require the conversion coefficient factor individually. The simpler measurement method for the density measurement becomes stringent. It is well known among the tunnel engineer that the suspended dust reflects flashlight during flashed photography. By using this phenomenon, the authors proposed the dust density measuring technique for suspended dust by the flashlighted image of digital camera. In this study, the authors propose the simple dust measuring technique by using artificial neural network. And the authors applied it in real tunnel construction site, and compare with the results obtained by the digital dust meter, and the effectiveness of this method was confirmed.

Journal

  • Doboku Gakkai Ronbunshuu G

    Doboku Gakkai Ronbunshuu G 66(4), 194-200, 2010

    Japan Society of Civil Engineers

Codes

Page Top