Robust and Precise Measurement Method of Vehicles and Motorcycles for Cooperative Driving Safety Support System with Combination of HOG-SVM Detection and Discriminative Pixel-Pair Feature Tracking

In this paper, a new vehicle measurement system with image processing for cooperative driving safety support system (DSSS) is proposed. Image sensors for cooperative DSSS are required to have extremely high performance under any condition changes or environmental noise. However, there have been no known image sensor systems to meet the requirements. Therefore, the authors developed their own image sensor system. Inputs to the system are gray-scale images from a roadside video camera. The system consists of a detector and a tracker. The detector is based on histogram of oriented gradient feature and support vector machine trained with a newly developed efficient training method. The tracker is based on discriminative pixel-pair feature. Since both a detector and a tracker are of high-performance, they enables to correct miss of detection, error of size, position, or both, and false positives. Moreover, since the inputs to the system are gray-scale images, existing roadside cameras can be easily converted to image sensors for cooperative DSSS. The results of the authors experiments shows that the proposed system detects each vehicle with less than 1.0% misses and one false positive per frame, and can be processed in 30ms per frame in average. In addition, the system is robust to condition changes or environmental noise.

  • Availability:
  • Supplemental Notes:
    • Abstract reprinted with permission from Intelligent Transportation Society of America.
  • Corporate Authors:

    ITS America

    1100 17th Street, NW, 12th Floor
    Washington, DC  United States  20036
  • Authors:
    • Ogiuchi, Yasuo
    • Higashikubo, Masakatsu
    • Ono, Yuuki
    • Kurita, Takio
    • Nishida, Kenji
    • Inayoshi, Hiroaki
    • Arata, Reiji
  • Conference:
  • Publication Date: 2011

Language

  • English

Media Info

  • Media Type: Digital/other
  • Features: Figures; Photos; References; Tables;
  • Pagination: 12p
  • Monograph Title: 18th ITS World Congress, Orlando, 2011. Proceedings

Subject/Index Terms

Filing Info

  • Accession Number: 01449839
  • Record Type: Publication
  • Files: TRIS
  • Created Date: Oct 22 2012 9:14AM