Access to SPIE eBooks is limited to subscribing institutions. Access is not available as part of an individual subscription. However, books can be purchased on SPIE.Org
Chapter 4:
Sampled Imager Design and Optimization
Published: 2000
DOI: 10.1117/3.353257.ch4
The way an image is processed and displayed is just as important as the blur and sampling characteristics of the sensor. In this chapter, the importance of interpolation and image reconstruction is demonstrated. The classical design rules for sampled imaging systems are reviewed. A new method for optimizing the design of sampled systems, the “MTF Squeeze” approach, is presented. Finally, three exercises are provided where MTF Squeeze is used to optimize the design of imaging systems. 4.1 INTERPOLATION AND IMAGE RECONSTRUCTION The sampling artifacts associated with out-of-band spurious response can be removed by the display or image reconstruction process. Multiple display pixels per sensor sample are used, and the sensor samples are interpolated to provide the intensity values for the added pixels. It is possible to remove essentially all of the out-of-band spurious response with little degradation to the transfer response of the sensor. That is, image interpolation can remove much of the bad without affecting the good; there does not need to be a degradation caused by the interpolation process. Interpolation requires some amount of image processing, and the display must have more than one pixel per sensor sample. So, there may be a system cost or complexity penalty for interpolation. Interpolation enhances the fidelity of the displayed image. From a spatial-domain viewpoint, the reconstructed image should match the pre-sampled image between samples. Large, rectangular or square display pixels have sharply defined edges; these visible edges represent spurious content which did not exist in the pre-sample image. On the other hand, large, Gaussian shaped display pixels tend to average the image between samples, blurring and softening the image. However, a mathematical interpolation can provide a good estimate of the pre-sample intensity pattern between the sensor samples. Multiple, smaller display pixels can then be used to generate the correct intensity pattern on the display. From a frequency-domain viewpoint, the display and any associated signal processing should provide a filter that discriminates against the spurious response while retaining the baseband transfer response.
Online access to SPIE eBooks is limited to subscribing institutions.

Image processing


Imaging systems

Image restoration

Image enhancement

Modulation transfer functions

Image interpolation

Back to Top