Four decades of optical information processing

Abstract

While the origins of optical information processing can be traced to the work of Ernst Abbe, active research in this area began in the 1950s. During this first decade, the idea of coherent Fourier plane filtering flowered, and applications to image processing were the initial focus of attention. Filtering power was limited by the difficulty of controlling both amplitude and phase transmission through the focal plane. The 1960s saw the introduction of the laser and the discovery of the interferometrically generated filter, which allowed implementation of much more sophisticated filters than had previously been possible. Emphasis increased on the processing of radar signals, and matched filter pattern recognition. The 1970s saw a major shift of attention to discrete data, and development of a multitude of techniques for processing vectors and matrices. The shortcomings of analog (as opposed to digital) processing became especially clear in this decade. In the 1980s attention turned to three subjects, often all included in the term optical computing: all optical digital processing, optical interconnects, and optical neural computing. These three areas remain active at the beginning of the 1990s. Some speculation will be provided regarding what emphasis we will see throughout the 1990s.