面向交通流信号分析的小波阈值去噪方法

Research on wavelet threshold denosing method for traffic flow signals analysis

为了提高交通流信号的去噪效果,提出了一种自适应逐层调整阈值和改进阈值函数的去噪算法。该算法结合自然对数函数的特性,使阈值随小波系数分解层数自适应的改变;加入了指数函数,构造一种介于硬、软阈值之间变化的新阈值函数,使去噪信号和原信号之间的误差减小。实验结果表明:相较于传统硬软阈值去噪,采用改进算法处理的信号直观效果好,且信噪比高和均方根误差小,运行效率高。改进后去噪算法各方面性能都得到了提升,具有较高的应用价值。

To improve the denoising effect in traffic flow signals,a denoising algorithm which adaptively adjusts the threshold and improves threshold function layer by layer was proposed.The algorithm combined the property of natural logarithm function and made the threshold change adaptively with the wavelet coefficient decomposition layer.Meanwhile,the exponential function was added to construct a new threshold function between hard and soft thresholds,which reduced the error between denoising signal and original signal.The experimental results showed that compared with the traditional hard and soft threshold denoising,the signal processed with the improved algorithm had better visual effect,with higher signal-to-noise ratio,lower root-mean-square error and higher operation efficiency.All properties of the denoising algorithm improved.Therefore,it owns a better practical application value.