The method design of coordinates mapping in PIV measurement system for flow field of impinging streams

Abstract

In the Particle Image Velocimetry (PIV) measurement system for flow field of impinging streams, in order to realize the calculation of real flow field data from sequence image, the method of coordinates mapping applying neural network is designed. This method is to put a plat with normal pitch grid in the area of transparent Liquid-continuous impinging streams reactor, with the grid plan going through the cylinder axis of the reactor filled with water. Digital camera is fixed outside the container, with the light axis vertical to the grid plan. Software tool is utilized to obtain Educate vector collect for educating neural network. The network which finished training can counterchange the pixel coordinate of the measured plan in the liquid into real coordinate.