Image registration and focus stacking on mobile platforms

摘要

A mobile device comprising a processor configured to merge a plurality of images each having a distinct focus position in a scene to generate a focus-stacked image that encompasses all focus positions of the plurality of images. Also, an apparatus comprising a camera configured to capture a plurality of color images for one scene each with a distinct focus distance and a processor configured to register the plurality of color images to generate a plurality of registered color images, generate a plurality of luminance (Y) channels from the plurality of registered color images, and stack the plurality of Y channels to generate a focus-stacked Y channel.

主权项

1. A mobile device comprising:
a memory; and a processor coupled to the memory, the processor configured to merge a plurality of images each having a pre-selected focus point in a scene to generate a focus-stacked image that encompasses all focus points of the plurality of images, wherein merging the plurality of images comprises decomposing the plurality of images via discrete wavelet transform (DWT) to generate a plurality of pyramidal structures each having an equal number of levels, wherein for each of the plurality of images, each level has four frequency bands denoted as low-low (LL), low-high (LH), high-low (HL), and high-high (HH) bands, and wherein for each of the plurality of images, the LL, LH, HL, and HH bands of a first level in the pyramidal structures represents the LL band of a second level immediately above the first level in the pyramidal structures.