摘要
针对高速运动条件下交通实时监测图像的失真和去噪问题,介绍小波阈值图像去噪过程和传统GCV阈值的原理,引入整数小波变换和合并递推运算的思想,提出一种快速递推GCV阈值的小波图像去噪方法。仿真实验分析了传统GCV阈值和快速递推GCV阈值去噪的效果,并比较了两种算法的复杂度。结果表明,快速递推GCV阈值算法不仅保留了传统GCV阈值算法去噪效果明显的特点,而且很好地解决了算法复杂度较高的问题,具有良好的应用前景。
Considering the image distortion and de-noising problems of traffic images from real-time monitoring on high speed moving condition,introduces the wavelet threshold de-noising process and the principles of traditional GCV threshold.An image de-noising method on fast recursive GCV threshold function of wavelet theory is proposed,which is based on integer wavelet transform and merge recursive operations.The de-noising effects between tradition threshold and fast recursive GCV threshold function is compared through the computer simulation,as well as the algorithm complexity features.The simulation results show that FR-GCV threshold function can remove the noise and retain the characteristic of the traditional GCV threshold function.Meanwhile,it can solute the complexity problem of tradition GCV threshold function and has a good application prospect.
出处
《微计算机信息》
2011年第3期194-196,共3页
Control & Automation
基金
基金申请人:马宏锋
项目名称:基于软计算的分布式智能交通视觉监控技术研究
基金颁发部门:甘肃省自然科学基金项目(096RJZA084)
基金申请人:党建武
项目名称:软计算技术在铁路智能分布监控系统中的应用研究
基金颁发部门:高等学校博士学科点专项科研基金项目(20060732002)
关键词
交通图像
小波分解与重构整数小波变换
GCV阈值函数
Traffic image
Wavelet decomposition and restructuring
Integral wavelet transform
GCV threshold function
