For the first time, physicists are using a statistical approach to process complex light patterns

Everybody uses digital cameras to take a picture of a particular subject or, once in a while, to take a selfie. When we take pictures, we are measuring the intensity of light from the surface of our subject.  There is, however, another property called the phase or wavefront of light which is much more sensitive to the state of the subject we are imaging.

Modern technologies are ubiquitous, for example, in the opto-electronic and biomedical research fields. Modern instrumentation systems require faster and more precise evaluation tools. Optical evaluation methods, particularly those that are wavefront-based are in great demand. Developing novel optical techniques, therefore, not only advances our understanding of light but also harnesses its practical benefits.

At the National Institute of Physics in UP Diliman, researchers led by Dr. Percival Almoro have developed a digital technique for measuring light as a means to investigate test objects. In particular, it can detect special light patterns obscured in noisy signals. These special light patterns are important in optical metrology – or the science of performing measurements with light – because they can be correlated to the minutest changes that happen to the light source or the test object through which light is scattered. With this method, optical metrologists will be able to evaluate different light patterns faster and very precisely, thereby saving time and money.

The pioneering technique, completely developed in UP, uses a statistical approach to process light patterns. From the 1980s up to the present, other researchers mainly concentrated on straightforward approaches such as image processing and mathematical transformations. The drawback of image processing is its sensitivity to noise while mathematical transformations apply only to simple light patterns. But the statistical approach, which has never been explored before, is robust against noise and can be used to evaluate complicated light patterns.

The statistical approach is more powerful in identifying with precision the location of a test object, in this case, a five-centavo coin (a). Unlike the “coarse” image of the hologram phase map, (b), the image of the standard deviation contrast map from the new method has high resolution and less noise (c). The bell-shaped curve reveals the precise location of the coin at z = 49.00 mm +/- 0.25 mm (d).

It was in 2011 when graduate student Anne Margarette Maallo had a serendipitous moment while working on a phase retrieval technique. She found that the standard deviation of phase images turned out to yield one-dimensional signature values. Subsequently, the research team developed the statistical concept into an algorithm to form a two-dimensional signature image called contrast map. The algorithm has been applied in the detection of optical whirlpools [1], evaluation of sub-surface defects [2], and measurement of temporal coherence of laser sources [3].

The present study extends its application to three-dimensional space, specifically, in the precise localization of rough test objects. In principle, the resolution of the technique, or the shortest distance between two objects that can still be distinguished, is less than 1 micron limited only by the wavelength of light.

Almoro will be presenting the ground-breaking technique in an invited plenary talk at the International Conference on Speckle Metrology on August 24-26, 2015 in Guanajuato, Mexico.

The research is supported by the UP Diliman Office of the Vice Chancellor for Research and Development, the National Research Council of the Philippines, and the UP System Office of the Vice President for Academic Affairs through its Enhanced Creative Work and Research Grant.