Effect of Sensor Geometry on the Performance of PIV Interrogation

* Final gross prices may vary according to local VAT.

Get Access

Abstract

This paper describes the mathematical investigation of the effect of the sensor geometry, i.e. the pixel size and pixel fill ratio, on the performance of PIV interrogation. Two sub-pixel estimators are investigated: the particle-image centroid, and the Gaussian peak fit. It is found that no bias errors occur when the particle-image diameter is at least two pixels, and the measurement error is determined by random errors only. When particle images are much smaller than one pixel, an irrecoverable signal loss deteriorates the measurement performance. For intermediate resolutions the bias errors are of the same magnitude as the random errors. It is demonstrated that image blurring by de-focussing reduces the bias error, but increases the random error. The analysis shows that sensors with a high fill ratio have a better performance.