人眼视觉特性与粗糙集结合的X射线图像增强算法

An Enhancement Method for X-Ray Image Using Rough Sets and Human Visual System

针对X射线图像的低亮度及噪声造成的图像对比度差和图像模糊的问题,提出了一种采用粗糙集理论和人眼视觉系统特性的图像增强算法(HREM).先将原始图像进行灰度拉伸,再结合人眼视觉系统特性,利用粗糙集理论将图像分成边缘细节图像和平滑图像,然后对边缘细节图像做指数变换增强,对平滑图像做直方图均衡化,最后将处理好的两子图进行重叠,同时消除重叠后图像的随机脉冲噪声.实验结果表明,HREM方法不仅能较好地保持图像的边缘细节信息,有效地增强原图像的对比度,而且消噪能力强,整体视觉效果好.

The X-ray image from direct digital radiography system is captured in non-uniform and low illumination conditions,and is blurred by noise. A new enhancing method is proposed on the basis of rough set theory and human visual system （HREM）. The grays of the original image are stretched, and then the stretched image is used to generate two new images （edge-detail image and smooth image） using the rough set theory and the human visual system. The edge-detail and smooth images are enhanced by the exponential transform and histogram equalization, respectively. Finally, the two enhanced images are overlapped into one image, and the random impulse noise in the overlapped image is reduced. Experimental results show that the HREM method can suppress noise while preserving edge and detail element, and enhance visual effect effectively.