| 摘要 |
When creating distance information for each of a plurality of objects in a plurality of images having the same viewpoint, the image pickup apparatus detects distances to the objects intermittently, and predicts distances to the objects of the other images using the detected distances. Object regions are extracted from each of the images, the extracted regions are compared between images to predict the correspondence relationships, and on the basis of the predicted correspondence relationships, the distance information of the regions of the images subjected to distance detection are assigned to the corresponding regions of the images to be subjected to distance prediction. By performing distance detection intermittently, lower power consumption is achieved, and by dividing the images to be subjected to distance prediction into regions on a per object basis and assigning distance information to each region, mismatching in the contours of the objects in the predicted distance information with the objects in the image is prevented from occurring. |
