|
Information processing plays an important role in science and technology. As we have seen throughout this book, optical systems can perform a wide variety of transformations on signals that can be displayed over 2D planes. This is in contrast to electronic processors, where signals are presented only as a one-dimensional temporal signal. This chapter deals with a few examples of harnessing optical methods to the processing of images and other signals.
An important field of optical image processing consists of pattern recognition, pattern localization and pattern tracking. Compared to digital methods, the atractivity of the optical correlator is rooted in its highly parallel operation and shift-invariance. This field was initiated by the work of VanderLugt [36] with the 4 f optical correlator as already discussed in chapter 5. However, the unavailability of proper optical-electronic-optical interfaces hindered the progress in this field until more advanced devices were fabricated. Unfortunately for optical scientists, the rapid advance of digital computers that occurred meanwhile did not leave much room for competition from optics. Nevertheless, there are various operations where the high parallelism of optics is still a significant advantage and there optical procedures can complement electronic processing rather than compete with it. The objective of this chapter is to introduce several procedures involving hybrid electro-optical signal processing. Specific applications to be addressed below are examples from which general procedures can be also deduced. In particular, some procedures and numerical algorithms will be reviewed that are useful for the design of DOEs. In addition to this main objective, this chapter provides several examples where the subjects studied in this book are put to work.
|
