A first image and a second image are obtained by sensing an image of a moving object from two different viewpoints using a first imaging device and a second imaging device. A plurality of feature points are detected from the first image. A disparity distribution representing a disparity for each feature point is obtained by performing stereo image processing using the first image and the second image. A threshold is determined for disparities in the disparity distribution. A feature point having a disparity exceeding the threshold in the disparity distribution is classified as an intra-moving-object feature point. An image area of the moving object in the first image is detected by using the intra-moving-object feature point.