高级检索

    基于自适应分形的射线图像增强算法

    The X-ray Image Enhancement Algorithm Based on an Adaptive Fractal Method

    • 摘要: 射线图像的主要特征是灰度对比度不强,人眼对其的视觉分辨或机器识别较为困难。利用图像各像素点的分形维数值作为加权值对原灰度图像进行增强,能够具有很高的灰度分辨率。但是其存在两个不足之处,一是对噪声非常敏感,二是其会使边缘粗化,因此提出了一种基于区域连通关系对H参数进行修正的新方法。该方法利用区域内中心点与周围灰度的连通关系,判断出其是噪声信号,背景信号或是边缘信号,分类对其分形维数进行修正,进一步对图像灰度值进行加权增强,取得较好的效果。仿真和试验结果证明,该方法既能保持分形对图像的高分辨率,又具有较好的抗噪性,适用于工业CT中的图像处理。

       

      Abstract: The main feature of X-ray images is of no significant gray contrast so that the flaw is difficult to identify. As the weighted value, the fractal dimension between the pixels is used to enhance the image, and can be of higher resolution. But there exists two shortcomings, one is very sensitive to noise, and the other is the roughening of the edges. A new method using the area connectivity to modify H parameter is proposed. The connectivity relationship between the center and ambient gray scale inside the region is used to distinguish noise signal, background signal, or the edge of the signal. By amending their fractal dimensions respectively, the image was further enhanced for better results. The simulation and experimental results showed that the method not only maintained high resolution of the fractal, but also was of better noise immunity and applicable for the industrial CT application.

       

    /

    返回文章
    返回