血管内超声图像的预处理

对血管内超声图像进行噪声滤波和边缘增强是对图像进一步处理分析的基础工作。针对采集到的强背景噪声的血管内超声图像,双边滤波利用像素间的空间距离和灰度值之差来控制滤波效果,即平滑了图像又保持了边缘。Retinex算法把图像分解为反射物体图像和入射光图像,在增强管壁边缘的同时对噪声影响小。实验结果表明,综合利用双边滤波和Retinex增强算法,在滤除噪声增强边缘有较满意的效果,对血管壁边缘检测有明显的改善。

Pretreatment of intravascular ultrasound images

Noise filtering and edge enhancement of intravascular ultrasound images are the basic work for further image processing and analysis.According to internal ultrasound images with strong background noise,bilateral filtering can control filtering effect by using the spatial distance in pixels and the difference between gray values,which smoothes the images as well as retains the edge.Retinex algorithm decomposes the images into reflection objects images and incident light images,which enhances the edge of blood vessel wall and has a little influence on noise.The experimental results show that bilateral filtering and Retinex enhancement have a satisfying effect on filtering noise and enhancing the edge of blood vessel wall,which improves the detection of the edge of blood vessel wall distinctly.