EAI Endorsed Transactions on Cognitive Communications 15(5): e2

Research Article

A Method for Detecting Damage of Traffic Marks by Half Celestial Camera Attached to Cars

Download12 downloads
  • @ARTICLE{10.4108/eai.22-7-2015.2260306,
        author={Takafumi Kawasaki and Takeshi Iwamoto and Michito Matsumoto and Takuro Yonezawa and Jin Nakazawa and Kazunori Takashio and Hideyuki Tokuda},
        title={A Method for Detecting Damage of Traffic Marks by Half Celestial Camera Attached to Cars},
        journal={EAI Endorsed Transactions on Cognitive Communications},
        volume={15},
        number={5},
        publisher={EAI},
        journal_a={COGCOM},
        year={2015},
        month={8},
        keywords={smart sensing, car, camera, image processing},
        doi={10.4108/eai.22-7-2015.2260306}
    }
    
  • Takafumi Kawasaki
    Takeshi Iwamoto
    Michito Matsumoto
    Takuro Yonezawa
    Jin Nakazawa
    Kazunori Takashio
    Hideyuki Tokuda
    Year: 2015
    A Method for Detecting Damage of Traffic Marks by Half Celestial Camera Attached to Cars
    COGCOM
    EAI
    DOI: 10.4108/eai.22-7-2015.2260306
Takafumi Kawasaki1,*, Takeshi Iwamoto2, Michito Matsumoto2, Takuro Yonezawa1, Jin Nakazawa1, Kazunori Takashio1, Hideyuki Tokuda1
  • 1: Keio University
  • 2: Toyama Prefectural University
*Contact email: drgnman@ht.sfc.keio.ac.jp

Abstract

Roads are becoming deterioration in everywhere. In some places, traffic marks painted on roads are damaged thus needed to be updated. Municipalities must manage road condition and traffic marks (road painting). It is the municipalities task to manage those roads using, for example, special inspection cars and human eyes. However, the management cost is high if a city contains many roads. This paper proposes a mechanism that automates this management. Our idea is to leverage cameras attached to garbage trucks, which run through the entire city almost everyday. The mechanism collects road images and detects damaged traffic marks using an image recognition algorithm. This paper shows the algorithm and reports the benchmark results. The benchmark showed that the mechanism can detect the damaged traffic marks with 76.6% precision.