血管内超声图像序列的自动降噪方法  

Automatic denoising method of Intravascular Ultrasound image sequence

作　　者：董路[1] 汪友生[1] 金铭[1] 

机构地区：[1]北京工业大学电子信息与控制工程学院,北京100124

出　　处：《电子测量技术》2015年第2期47-50,共4页Electronic Measurement Technology

摘　　要：提出一种血管内超声图像序列的自动降噪方法。首先对图像序列中每帧图像进行中值滤波,抑制图像脉冲干扰及斑点噪声,保护边缘少受模糊,然后采用图像序列均值滤波方法抑制图像加性随机噪声,提高图像信噪比,并利用小波软阈值降噪的多尺度分析特性,对均值图像的细节加以分辨,再结合算术运算得到最终降噪效果。对比中值滤波、小波软阈值降噪等传统滤波方法,提方法能得到更好的降噪效果;利用所提方法对连续不同帧数血管内超声图像进行降噪处理,当图像帧数为10时,图像序列的降噪效果更好;通过峰值信噪比、降噪时间和边缘检测效果的对比,可以发现本文所提方法能改善降噪效果、提高边缘检测的准确度,保证降噪效率。To propose an automatic denoising method of Intravascular Ultrasound（IVUS）image sequence.In the process of the proposed method,each frame of image sequence is denoised with median filtering,which not only suppresses pulse interference and speckle noise,but also protects edge from blurring.After that,the mean of image sequence is calculated to restrain additive random noise and improve the image signal-to-noise ratio,and then the detail of the mean image is distinguished based on wavelet soft-threshold noise reduction method which has multiscale analysis property.Moreover,arithmetic operations also need to be performed to obtain finally results.Compared with traditional denoising method such as median filtering,wavelet soft-threshold denoising method,the proposed method has better noise reduction effects.Different number adjacent frames are used to verify the effectiveness of the proposed method,when the frames number is 10,the noise reduction effect is much better than other numbers.Comparison of denoising time,PSNR,and edge examination effect shows that the proposed method can improve the denoising effectiveness and the accuracy of edge detection and ensure the efficiency.

关 键 词：血管内超声 图像处理 自动降噪 综合滤波 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]