Hourly and Spatial Analysis of Degrees of Congestion in the Observation Camera Scene by Accumulating Local Changes in the Image Sequence

Abstract

We can see a lot of observation cameras at public or private facilities such as stations, halls, shopping malls, hallways and gates. Some janitors of the facilities want to detect and investigate passing people or congestion existences in the camera scene. In order to respond to these requests, we extracted moving areas in the image and accumulated these data and then displayed them on the monitor. Consequently, we obtained spatial distributions and hourly changes of congestion in the observation camera scene. To extract moving areas, we estimated the differences of correlation value of segmented small windows in the image. This paper describes our procedure to analyze the congestion from the image sequence of observation camera, and some experimental results.