PARTICLE DISPLACEMENT TRACKING TECHNIQUE AND CRAMER-RAO LOWER-BOUND ERROR IN CENTROID ESTIMATES FROM CCD IMAGERY

An efficient video based, multi-frame particle displacement tracking t echnique is presented. The technique is demonstrated on a thermocapill ary flow. The Cramer-Rao lower bound centroid measurement uncertainty is derived for Gaussian particle images with Poisson noise recorded on a CCD array. The optimal particle image diameter is determined to be approximately 1 1/2 pixels. The centroid error increases for particle image diameters larger or smaller than the optimal value. The particle centroid results are carried over to the case of a cross-correlation peak centroid estimate. Particle tracking techniques are shown to have higher accuracy than auto or cross-correlation techniques when the in dividual velocity measurements are averaged over a comparable correlat ion subregion.