Scatterometry Using Speckle Correlations

<p dir="ltr">Material acquisition is the process of inferring material properties from observations. We focus on estimating the scattering phase function–the angular distribution of scattered light—which affects translucency and encodes information about composition. Accurate phase functions enable applications in computer graphics, medical imaging, and materials science. </p><p dir="ltr">Existing approaches either reduce optical density by dilution (restricting them to liquids) or invert the radiative transfer equation with differentiable rendering (susceptible to ambiguities from a non-convex inverse problem). To address the shortcomings of this prior work, this work introduces an approach that uses speckle correlation measurements of material samples, which can be mapped to phase function values using analytical expressions. This approach has been recently demonstrated using acquisition systems that limit the scanning range to angles less than 4◦ . </p><p dir="ltr">We introduce a computational imaging system that acquires phase functions over angular ranges approaching 180◦ . The system applies to solids and liquids and to optically thin or thick samples. Two mutually coherent continuous-wave laser beams, fixed to a goniometer with a close angular separation, illuminate the sample. A camera records the corresponding speckle patterns; the phase function along the bisector of the two beams is proportional to the measured speckle correlation. By rotating the illuminators about the sample, we measure the phase function over a wide range without the angular limitations of 4f imaging. The results support graphics (e.g., photorealistic augmented reality) and broader domains including noninvasive medical diagnostics, remote sensing, and quality-assurance particle sizing.</p>