Abstract

A two-dimensional sampled image can be represented by an equal number of spatial frequencies in the Fourier domain. However, due to physical limitations of the sampling process, there exists an effective cutoff frequency beyond which no information is preserved. The knowledge of this cutoff frequency is very important, since a considerable amount of additional noise can be introduced by frequency components above this cutoff. Availability of a sharp edge within the image allows, within the linear theory, the estimation of the transfer function of the digitizing process itself. This calculated transfer function may be used for image enhancement below the cutoff frequency. In addition, significant amount of data compression may be achieved by removing all spatial frequency terms above the cutoff frequency. An example of this technique is developed utilizing a tribar resolution chart sampled at 1024 × 1024 points.

© 1972 Optical Society of America

X-Ray Image Enhancement by Least-Mean-Square Estimation

David G. Cain
Appl. Opt. 11(12) 2949-2955 (1972)

Procedure for Over-all Lens Evaluation Using OTF Data

David Dutton
Appl. Opt. 11(5) 1091-1100 (1972)

Dynamic range compression and detail enhancement algorithm for infrared image

Gang Sun, Songlin Liu, Weihua Wang, and Zengping Chen
Appl. Opt. 53(26) 6013-6029 (2014)

Reusable Optical Image Storage and Processing Device

J. Feinleib and D. S. Oliver
Appl. Opt. 11(12) 2752-2759 (1972)

Detection of Weak Optical Images with Photon Counting Techniques

John Amoss and Frederic Davidson
Appl. Opt. 11(8) 1793-1800 (1972)