Automated Multimodal Diagnostic System of Social Infrastructures by Inspection Experts' Skill Extraction

2017 Fiscal Year Final Research Report

• Project/Area Number: 16H06680
• Research Category: Grant-in-Aid for Research Activity Start-up
• Allocation Type: Single-year Grants
• Research Field: Perceptual information processing
• Research Institution: The University of Tokyo
• Principal Investigator: FUJII Hiromitsu 東京大学, 大学院工学系研究科(工学部), 特任講師 (30781215)
• Research Collaborator: ASAMA Hajime 東京大学, 大学院工学系研究科, 教授 (50184156) ; YAMASHITA Atsushi 東京大学, 大学院工学系研究科, 准教授 (30334957) ; Im Jonghoon 東京大学, 大学院工学系研究科 ; Kasahara Jun Younes Louhi 東京大学, 大学院工学系研究科 ; YANAGIHARA Yoshitaka 東急建設株式会社 ; NAKAMURA Satoru 東急建設株式会社 ; TAKAHASHI Yusuke 東急建設株式会社
• Project Period (FY): 2016-08-26 – 2018-03-31
• Keywords: インフラ点検 / 自動診断 / センサ情報処理

Outline of Final Research Achievements

Aged deterioration of social infrastructure is rapidly getting serious. At conventional inspection sites, hammering tests have been carried out by skilled inspectors. At present, for safety and security, highly reliable automatic diagnostic technologies which replace hammering tests by the skilled inspectors are urgently required.
In this research, in order to realize high efficiency, high accuracy of diagnosis, and robustness against difference inspection targets in different environments, we aim to construct an automatic diagnosis system using multimodal signal. We worked on the following two approaches, and achieved the system construction and verification by experiments. 1) Improvement of robustness against differences in inspection targets and environments by using knowledge and working protocol of skilled inspectors. 2) Detection of crack direction against surface of the concrete using multimodal information obtained from acoustic sensors and a visual sensor with a laser scanner.

Free Research Field

音響など多種センサ信号と機械学習を用いた社会インフラの自動診断，およびコンピュータビジョン