Neural network PTV using a self-organizing maps method

Abstract

The neural network technique using a self-organizing maps (SOM) model is used for the particle tracking algorithm in PIV measurement of fluid flows. The new algorithm is based on that of La Bonté [2] with further improvement for larger dynamic range of velocity and more robustness against unpaired particle images between two frames. Tests are made with typical synthetic images (PIV-STD) as well as some experimental images.