Statistical assessment of the inner shape of a channel based on laser speckle contrast imaging

In this work, we employed laser speckle contrast imaging coupled with the analysis of several statistical moments on the recorded images to recognize the shape of an object through the analysis of the pattern of a flow around it. We postulated that a flow through different geometries would generate a unique flow structure and thus the mathematical analysis of the differences can be utilized to detect and recognize various shape geometries. To this aim, the flow of scattered liquid that had passed through objects of different shapes was illuminated by a laser beam and the diffused speckle image was recorded by a camera. In the computer, the captured speckle image was first converted to a flow image and then mathematically analyzed to recognize the shape of the objects. Three out of the eight statistical metrics were found to be the best candidates for the recognition of shapes, thus proving our hypothesis.