Sub-pixel shift estimation of image based on the least squares approximation in phase region

This paper proposes a novel method to estimate noninteger shift of images based on least squares approximation in the phase region. Conventional methods based on Phase Only Correlation (POC) take correlation between a image and its shifted image, and then estimate the non-integer shift by fitting the model equation. The problem with using POC is that the true peak of the POC function may not match the estimated peak of the fitted model equation. This causes error in non-integer shift estimation. By calculating directly in the phase region, the proposed method allows the estimation of decimal shift through least squares approximation. Also by utilizing the characteristics of the natural image, the proposed method limits adoption range for least squares approximation. By these improvements, the proposed method improves the estimation and achieves high accuracy.