Autofocus method using similarity measure based on inter-color channel correlation

Abstract

Recently, an autofocus method using dual aperture and color filters has been shown to be fast and cost-effective, and thus suitable for compact digital cameras. The dual aperture separates the incoming light rays and the color filters on the apertures transmit only a single color component. Hence, each light ray passing through an aperture can be separately captured by the corresponding color sensors. The captured monochromatic images of two different color channels are spatially shifted with respect to each other in propotion to the defocus. The disparity between the different color channel can be converted to the distance for the focusing lens to move. However, since the disparity between two different color images cannot well be measured by the conventional methods, we propose a new similarity measure between the dichromatic channel images (e.g. R and B color channels). Experimental results show that the proposed measure is more effective than the existing measures especially when the region of focus measuring is small. This advantage enables to focus the small object and reduce the processing time.