Accurate measurement of translational shifts by adaptively masking phase correlation

The measurement of translational shifts has an impact on applications in medical imaging, remote sensing and structural health monitoring. Phase correlation (PC) is a well-known method to measure displacement. However, the PC matrix is often degraded by phase noise and aliasing. Presented is a robust PC method based on adaptive masking and essentially non-oscillatory smoothing to reduce phase noise and aliasing. With this approach, translational shifts can be measured with an accuracy of 1/300 of a pixel as shown in the experimental results.