Study on automatic chalk marks recognition for productivity enhancement of infrastructure inspection

2021 Fiscal Year Final Research Report

• Project/Area Number: 19K04580
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 22020:Structure engineering and earthquake engineering-related
• Research Institution: Yamaguchi University
• Principal Investigator: Kawamura Kei 山口大学, 大学院創成科学研究科, 教授 (70397991)
• Project Period (FY): 2019-04-01 – 2022-03-31
• Keywords: 社会基盤施設 / 維持管理 / 点検 / データ管理 / 深層学習 / オープンデータ

Outline of Final Research Achievements

Current methods of practice for inspection of concrete tunnel typically involve visual assessments and drawing notes conducted manually by trained inspectors. The labor intensive and time-consuming natures of manual inspection have engendered research into development of method for automated defect identification using camera and image processing techniques. This study proposed an automatic chalk marks recognition by using deep learning. Then, in Japan, periodic inspection of infrastructure is compulsory once in 5 years. Consequently, after infrastructure inspection, many valuable data sets are accumulated. However, the data are managed based on Excel format. That makes it difficult and inefficient to extract information from data. Also, incorrect data input is occurred, because data are input manually by a person. Therefore, the study developed the prototype of the open date system for infrastructure maintenance and management.

Free Research Field

維持管理工学

Academic Significance and Societal Importance of the Research Achievements

本研究は、現在の定期点検の生産性向上において、最もボトルネックとなっている課題を扱っており、点検成果物の作成時間を大幅に短縮する基盤技術に関する研究を行った。さらに、定期点検結果など道路管理機関に蓄積されるデータのオープン化システムを提案またプロトタイプシステムを開発しており、今後のインフラメンテナンス工学の発展のために大きく貢献できる。