声场干涉条纹图像的降噪研究与性能仿真

摘要

Sound field interference fringe image generated in target localization algorithm using a guide source (GTL algorithm) is always accompanied with noise. Localization performance have been affected by noise quietly. In order to achieve the distance accurate location in the low signal to noise (SNR), wavelet thresholding denoising is proposed. Because of the hard and soft thresholding function to remove noise of signals have disadvantages, a new method of thresholding function is presented. Denoising performance can be efficiently improved by binary image setting after wavelet denoising. The simulation results show that the distance location efficiency of the GTL algorithm is improved by the denoising processing of using the wavelet method.%引导源目标定位算法形成的声场干涉条纹图像中常附带有噪声,这对定位性能的影响非常大.为了实现低信噪比条件下对目标的准确定位,根据算法仿真结果的特点,提出了利用小波阈值去嗓法对图像进行降噪处理.针对软硬阈值去噪的不足,采用了阈值函数的改进方案.对小波去噪后图像再进行二值化设置,能有效提高算法去噪效果.仿真结果表明小波降噪处理能有效提高引导源目标定位算法在低信噪比条件下的定位性能.