Abstract
In this paper, a multiple-layer neural network model has been applied to the Fourier transform method in 2-D Particle Image Velocimetry to improve the measurement accuracy. The input information of the neural network is the complex phase that is extracted from the Fourier transforms of two images captured in a short time interval, and the output is the spatial shift of the pattern on the images. The learning is performed by a conventional error back propagation method. The performance test shows that the present method is robust against velocity fluctuation and the computing time can be reduced to about 75% of that of the original Fourier transform method.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have