摘要 |
A pair of images with different capture times are obtained during live-view image or video shooting (step S10). Motion vectors are computed from the pair of images, and moving object detection based on the motion vectors is carried out (steps S12, S14). Also, a difference image is generated from the pair of images, and moving object detection based on inter-frame differences is carried out (steps S16, S18). The two types of moving object detection results are combined in order to detect a moving object (step S20). Consequently, this enables a fast-moving object, which cannot be detected using the motion vectors, to be detected while maintaining the accuracy of slow-moving object detection. |