Optical interferometry is a full-field, noncontact, high-sensitivity technique for precision measurement. Optical interferometry outputs a fringe pattern or a sequence of fringe patterns. Some noninterferometric measurement techniques also have fringe patterns as their outputs. In this book, we explore methods for analyzing fringe patterns to extract phase distributions—a process called phase extraction. The extracted phase distributions are then converted to physical quantities being measured—this process is called phase conversion. Phase conversion depends on the configuration and calibration of optical measurement systems. It is only briefly discussed in this book because of the wealth of information provided on the topic in the references in Section 1.1. We frequently convert phase of 2p to one wavelength, merely for the sake of intuitiveness.
Fringe pattern analysis algorithms are expected to be accurate, automatic, and accelerated (A3). Accuracy is the rule of thumb for precision measurement. Accuracy and precision indicate the bias and variance of phase estimation, respectively. For simplicity, we use accuracy loosely to refer to both terms. Automaticity is important for an algorithm when it is integrated into a measurement system. The algorithm is expected to be able to process the fringe patterns with little human interaction, but with a high success rate nonetheless. Fast realization, either through algorithm optimization, or hardware acceleration, is also key to achieving “what you see is what you measure,” which is the goal in real measurement. Without good algorithms, measurement techniques that use fringe patterns for data representation will be restricted.1 Effective and efficient fringe pattern analysis has been enthusiastically pursued. As experiences in the field accumulate and computing techniques continue to develop, fringe pattern analysis techniques are advancing.
Online access to SPIE eBooks is limited to subscribing institutions.