基于曲波变换的接触网磨耗图像增强研究  

Research on Catenary Abrasion Image Enhancement Based on Curvelet Transform

作　　者：汪芳莉[1] 顾桂梅[1] 

机构地区：[1]兰州交通大学自动化与电气工程学院,甘肃兰州730070

出　　处：《铁道标准设计》2014年第2期112-116,共5页Railway Standard Design

摘　　要：针对采集到的接触线磨耗图像存在边缘模糊、对比度差等不足,将曲波变换应用于接触导线磨损图像的增强。首先对接触线磨损图像进行曲波变换得到不同尺度系数,其次在低频域利用分数阶微分能增强信号的中频成分、对信号的低频成分非线性保留的特点,对曲波域低频信号的纹理信息进行增强;同时在曲波域的高频域,利用曲波系数间尺度相关性,对高频系数的边缘和噪声进行分类,对分类后的信息进行边缘增强和噪声去除的处理。实验结果表明,所采用的算法和其他算法相比,主观和客观上都有明显改善。In this research, considering that the different shortcomings usually accompany the contact wire abrasion images collected, such as edge blurring, poor contrast ratio and so on, the curvelet transform technology was employed to enhance the abrasion image of contact wire. First, the contact wire abrasion image was processed through curvelet transform so as to get the coefficients of different scales. Next, by utilizing the characteristics that： the fractional differential can enhance the mid-frequency component of signal and can nonlinearly retain the low-frequency component of signal, the texture information of the low-frequency signal of the curvelet domain was frequency domain of the curvelet domain, the edge of the high-fr enhanced. Meanwhile in the equency coefficients and noise were classified by utilizing the scale correlation among the curvelet coefficients; and after classifying the information, the edge was enhanced and the noise was removed. Experimental results show that by comparison with other algorithms, this algorithm proposed in this paper has remarkably improved the subjective and objective effects.

关 键 词：电气化铁道 接触线磨耗 图像增强 曲波变换 分数阶微分 

分 类 号：U225.6[交通运输工程—道路与铁道工程] TP391[自动化与计算机技术—计算机应用技术]