摘要
提出一种改进多阈值小波包的去噪算法,解决了单一阈值对噪声去除不完全和对一些有用信号无差别去除的问题。应用在智能交通的图像去噪中,解决了不完全及错误去除图像信息的问题。首先采用小波包分解重构算法对图像进行预处理,得到更多的边缘细节。然后针对不同能量对应不同频段的特点,自适应地合理设置阈值,对不同频段下的噪声采用不同阈值去除。实验表明,该方法有效去除噪声,保留了图像的边缘和细节。
An improved multiple threshold wavelet packet de-noising algorithm is proposed, solving the problems of eliminating the noise incompletely and removing some useful signals without distinction. It was applied to the intelligent traffic image and got rid of the image signal de-noising incompletely and wrongly. First of all, the image was preprocessed using the decomposition reconstruction＇s algorithm of wavelet packet, and got more edge details. Then corresponding to different frequencies, threshold was set reasonable according to the characteristics of different energy adaptively, and different threshold was used to remove noise under different frequencies. Experiments showed that the method could remove the single noise effectively while preserving the image edges and details.
出处
《计量学报》
CSCD
北大核心
2016年第2期205-208,共4页
Acta Metrologica Sinica
基金
国家自然科学基金(61473248)
河北省自然科学基金(F2015203413)
河北省高等学校科学技术研究重点项目(ZD2014100)
关键词
计量学
去噪
小波包
多阈值
图像预处理
智能交通
metrology
de-nosing
wavelet packet
multiple threshold
image preprocess
intelligent traffic
