A Novel Inverse Imaging Approach to Brightfield Micro-PIV

Brightfield micro-PIV, which uses a collimated light source aligned with the optical axis of the imaging camera to illuminate particles throughout a seeded volume, is commonly employed to study various micro-flows. In micro-PIV, a microscope objective provides a narrow depth-of-view, isolating in-focus particles and defining the plane of velocity measurements. Particle image morphology depends on its position with respect to the focal plane, with particles blurring outside of this plane. Out-of-focus particles can be eliminated via image processing (e.g. intensity thresholding), but precise differentiation of in- and out-of-focus particles is difficult, resulting in lost information and degraded vector fields. Here we present a novel approach to processing brightfield micro-PIV images based on inverse imaging. This approach implements a computational algorithm to extract in-focus particles and suppress noise in PIV images by solving the sparsity-regularized inverse problem that arises when a Gaussian function is used to model each particle’s intensity morphology. We show application of this approach using measurements of the flow generated by a tiny, free-flying insect and demonstrate enhanced image quality and vector field quality.