论文部分内容阅读
数字粒子图象测速技术 ( DPIV- Digital Particle- Imaging Velocimetry)已在国内外得到广泛的重视和应用 ,但目前其最大的问题是精度问题 .由于 DPIV的图象数据是用 CCD摄像机经相应的图象卡采集示踪粒子图象得到的 ,这样在实验过程中不可避免引入的噪声 (主要是示踪粒子大小、示踪粒子数量、诊断窗口大小、诊断窗口内的速度梯度和量化效果等引入的噪声 )降低了实验测量的精度 .本文应用小波变换的多分辨率特性 ,对 DPIV图象 (模拟和实际图象 )进行去噪处理 ,并与维纳去噪和中值去噪进行比较 .比较结果发现 ,小波变换能提高 DPIV测量的精度 ,因而 DPIV图象只有通过小波去噪后再进行基于互相关算法重建速度场的计算才是最精确的 .
Digital Particle-Velocimetry (DPI) has been widely used both at home and abroad, but the biggest problem nowadays is its accuracy.Due to the fact that the DPIV’s image data is captured by the CCD camera Image card acquisition tracer particle images, so in the course of the experiment is inevitable to introduce noise (mainly tracer particle size, the number of tracer particles, the diagnostic window size, the diagnostic window of the speed gradient and quantization effects introduced Noise) reduces the accuracy of experimental measurement.In this paper, the multi-resolution characteristics of wavelet transform are used to denoise the DPIV images (analog and actual images) and compared with Wiener denoising and median de-noising. The results show that the wavelet transform can improve the accuracy of DPIV measurement. Therefore, the DPIV image is the most accurate calculation of the velocity field reconstructed by cross-correlation based on wavelet denoising.