Stereo-PIV using self-calibration on particle images

TLDR: In this paper, a stereo-PIV calibration procedure is developed based on fitting a camera pinhole model to the two cameras using single or multiple views of a 3D calibration plate, and a disparity vector map is computed on real particle images by cross-correlation of the images from cameras 1 and 2 to determine if the calibration plate coincides with the light sheet.

Abstract: A stereo-PIV (stereo particle image velocimetry) calibration procedure has been developed based on fitting a camera pinhole model to the two cameras using single or multiple views of a 3D calibration plate. A disparity vector map is computed on the real particle images by cross-correlation of the images from cameras 1 and 2 to determine if the calibration plate coincides with the light sheet. From the disparity vectors, the true position of the light sheet in space is fitted and the mapping functions are corrected accordingly. It is shown that it is possible to derive accurate mapping functions, even if the calibration plate is quite far away from the light sheet, making the calibration procedure much easier. A modified 3-media camera pinhole model has been implemented to account for index-of-refraction changes along the optical path. It is then possible to calibrate outside closed flow cells and self-calibrate onto the recordings. This method allows stereo-PIV measurements to be taken inside closed measurement volumes, which was not previously possible. From the computed correlation maps, the position and thickness of the two laser light sheets can be derived to determine the thickness, degree of overlap and the flatness of the two sheets.