Abstract
Measurement error of particle image velocimetry (PIV) associated with calibration functions based on polynomial camera model and pinhole camera model was evaluated numerically. The fitting error of each calibration function was evaluated using a virtual calibration grid generated by ray-tracing method. Pinhole camera model gave smaller calibration error than polynomial model. Error including grid-detecting procedure was also evaluated by the simulated calibration grid image. The error was larger than the fitting error only, and has no obvious dependency on the camera model. Stereo-PIV measurement was performed using the simulated particle image of the uniform flow field with the above calibration functions. Both camera models gave approximately the same magnitude of errors, but strongly sensitive to the angle between camera axis.
