To overcome a problem such that extracting individual vehicle correctly is difficult when a plurality of vehicles are in contact with each other via a shadow (make one combined area).

A background image generating means 2 generates a background image from an inputted picture, and a background difference image generating means 3 generates the difference image between the background image and the inputted image. A moving body detection means 4 detects a moving body area from a background difference image, and a shadow area separation means 5 detects a dividing object area from the moving body area based on a binarized pixel value to calculate the edge information of each pixel from the value of each pixel and that of a neighboring pixel in the moving body area. A moving body following means 6 follows the moving body based on a detected matter area with each dividing object area as the matter area or shadow area, based on the edge information.