GENERARTING 3D IMAGES USING MULTI-RESOLUTION CAMERA SET

摘要

Techniques for generating 3D images using multi-resolution camera set are described. In one example embodiment, the method includes, disposing a set of multi-resolution cameras including a central camera, having a first resolution, and one or more multiple camera groups, having one or more resolutions that are different from the first resolution, that are positioned substantially surrounding the central camera, Images are then captured using the multi-resolution camera set. A low-resolution depth map is then generated by down scaling the captured higher resolution image to lower resolution image. Captured lower resolution images are then up-scaled. A high-resolution depth map is then generated using the captured image of the central camera, the up-scaled captured images of the one or more multiple camera groups, and the generated low-resolution depth map. The 3D image of the captured image is then generated using the generated high-resolution depth map and the captured images.

主权项

1. A method for generating a three-dimensional (3D) image using a multi-resolution camera set, comprising:
disposing a set of multi-resolution cameras including a central camera, having a first resolution, and multiple camera groups, having one or more resolutions that are lower than first resolution and disposed substantially around the central camera; capturing images using the set of multi-resolution cameras; down-scaling the resolution of the captured image of the central camera to be substantially equal to the one of one or more resolutions associated with the cameras of the multiple camera, groups; generating a low-resolution depth map using down-scaled captured image of the central camera and the captured images associated with the multiple camera groups using relatively small depths; refining, the generated low-resolution depth map; upscaling the resolution of the captured images of the multiple camera groups to be substantially equal to the resolution of the captured image of the central camera; generating a higher resolution depth map using the captured image of the central camera, the up-scaled captured images of the multiple camera groups, and the refined generated low-resolution depth map; and generating the 3D image of the captured image using, the generated high-resolution depth map and the captured images.