Image Fusion for Night-time Pedestrian Recognition Assistance

Abstract

Fusion images have been generated by combining near infrared images of road areas and far infrared images in order to reduce vehicle-pedestrian accidents at night. In past studies, fusion images were generated by combining the entire area of near infrared images and far infrared images, which resulted in image information unnecessary for pedestrian recognition being fused, such as street illumination and the headlights of oncoming vehicles. It might be achievable to enhance efficiency in providing visual driving assistance by reducing these influences. The current research applied gradation processing to combine only the road area of near infrared images and far infrared images, and generate partial fusion images that primarily provide pedestrian and road area information. The effectiveness of this type of fusion image in providing visual assistance was compared to that of other image types in a volunteer test. Increased accuracy in identifying the location of pedestrians was shown when the new images were employed.

References

  1. (1) Traffic accident statistics annual report 2003, Institute for Traffic Accident Research and Data Analysis,(2004)
  2. (2) Tsuji, T., Hattori, H., Watanabe, M., Nagaoka, N.: Development of Night Vision System, IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS, Vol. 3, No. 3, p. 203-209(2002)
  3. (3) Suard, F., Rakotomamonjy, A., Bensrhair, A., Broggi, A.: Pedestrian Detection using Infrared images and Histograms of Oriented Gradients, IEEE IV2006, p. 206-212(2006)
  4. (4) Toyofuku, K., Iwata, Y., Hagisato, Y., Kumasaka, T.: The “Night View System” Using Near-Infrared Light, SAE Paper, 2003-01-0018(2003)
  5. (5) Herrington, W. F., Horn, B. K. P., Masaki, I.: Application of the Discrete Haar Wavelet Transform to Image Fusion for Nighttime driving, IEEE IV2005, p. 273-277(2005)

Author (organization or company)

Publication Information

Article of Honda R&D Technical Review Vol.19 No.2 (129 - 135)

PDF

PDF 397KB

Citation Information

Junji EGUCHI, Izumi KONDO, Hiroyuki KOIKE.
Image Fusion for Night-time Pedestrian Recognition Assistance
Honda R&D Technical Review Vol.19 No.2 pp.(129 - 135)2007.
Copy citation information
Copied