Application of phase unwrapping to image restoration

The restoration of phase information in two image processing problems is studied. First, two methods due to Knox and McGlamery for obtaining diffraction limited phase information of objects viewed through randomly fluctuating media are improved and evaluated on computer-simulated data. Although both methods yield good results it is concluded that Knox's algorithm is superior. Second, two techniques for acquiring the phase of a blur function given only the degraded image are evaluated. An existing method is improved and a new method is introduced. Experimental results indicate that these algorithms are sensitive to both the image class and edge effects.