Choice of Wavelets for Piv Image Compression

In this paper the wavelet-based image compression technique is applied to PIV processing for reducing the physical storage. To determine the effect of the choice of the wavelet bases, the standard PIV images are compressed by some known wavelet families, Daubechies, Coifman and baylkin families. It was found that high order wavelet bases provides good compression performance for compressing PIV Images, because they have good frequency localization that in turn increases the energy compaction. The reconstructed PIV image with lower compression ratio may emphasize particle edges at a relatively high spatial resolution, and the reconstructed PIV image with higher compression ratio may display the large-scale motion of particles and may deduce noisy. In this study, higher compression ratio, from 25% to 1.56%, can be realized without losing significant flow information in PIV processing. It can say that the wavelet image compression technique is effective in PIV system. INTRODUCTION It is well known that Particle Image Velocimetry (PIV) is now firmly established as a powerful fluid dynamics tool to measure instantaneous full-field velocity. Since digital images used in PIV technique are inherently voluminous, efficient data compression techniques are becoming an important problem for economy in storing and processing with high speed. The International Standard Organization (ISO) has proposed the JPEG standard for still image compression and MPEG standards for video compression. These standards employ discrete cosine transform (DCT) to reduce the spatial redundancy present in the images or video frames. We note that DCT has the drawbacks of blocking artifacts, mosquito noise and aliasing distortions at high compression ratios. However, the method of image compression that was often used in PIV is only to eliminate the low intensity pixels of image file. Because the low intensity pixels contribute little information about particle displacement, this type of image compression has very little effect on the accuracy of PIV. Recently, Hart (1998) employed sparse array image correlation to realize compression ratios of 30:1 or greater and high processing speed of PIV. Over the past decade discrete wavelet transform (DWT) has emerged as a popular technique for image compression application. DWT has high decorrelation and energy compaction efficiency. A wide variety of wavelet-based image compression schemes have been reported in the literature, however, few applications can be found in the area of fluid mechanics. Our motive of this study is to develop an application of wavelet technique to PIV image compression for minimizing the physical storage. It is a well-known fact that PIV image contains the redundancy of information, and this redundancy allows us to compress the image for economy in storing, transmitting or further processing. In this paper, we employ the wavelet-based image compression technique to the standard PIV images that was developed by The Visualization Society of Japan to realize 1 Copyright © 1999 by ASME higher compression ratio. To determine the effect of the choice of the wavelet bases, we compare the compression performance of wavelets of various orders within the same families. We then compare the performance of some known wavelet families, such as Daubechies, Coifman and baylkin bases. Finally, we discuss the optimal wavelet bases for PIV image compression. WAVELET COMPRESSION Let us consider a two-dimensional scalar field ( ) x f & and isotropic mother wavelet ( ) x& ψ by treating ( ) 2 1, x x x = & as vector. The family of wavelets ( ) x a b &