Driving video recorders are recently in widespread use. What is seen and what happens along the way can be recorded in videos. However, searching for a specific scene or event among such video collections is laborious. Hence, we propose a speed-adaptive street view image generation technique for the videos recorded by a driving video recorder. Visual summaries of the street scenes along the way can be provided, allowing users to retrieve a video clip corresponding to a specific road section very efficiently. The processing speed can be up to 400 fps for the image resolution of 640´360.
Fig. (a) Original frame captured by a driving video recorder. (b) Motion analysis map (generated from the driving video) for estimating the frame speed. (c) Generated street view image, enabling a user to quickly browse the street view passed. The user can also click the street view image to retrieve the video clip corresponding to a specific road section.