能强化细节的低剂量CR图像增强算法  

Low-dose CR image enhancement algorithm

作　　者：童莹[1] 

机构地区：[1]南京工程学院通信工程学院,江苏南京211167

出　　处：《计算机工程与设计》2010年第13期3132-3135,共4页Computer Engineering and Design

基　　金：江苏省自然科学基础研究计划基金项目(BK2009351);南京工程学院重点课题基金项目(KXJ07014);引进人才课题基金项目(KXJ07-67)

摘　　要：针对低剂量CR图像比度低、骨裂纹细节不清晰等缺点,提出一种基于自适应直方图均衡和改进小波变换的有效图像增强算法。采用自适应直方图均衡提高整幅图像的对比度,利用改进小波变换对骨裂纹等细节进行微调增强。实验结果表明,该算法不仅有效提升了整幅图像的灰度阶次,而且对细节信息也能有效保留,处理后的CR图像细节丰富,具有良好的视觉效果,PNSR和有效灰度级数进一步证明了它的有效性。Aimed at the disadvantages of low-dose CR images,a new image enhancement algorithm is put forward based on adaptive histogram equalization and wavelet transformation is improved.Firstly,the contrast of CR image is globally enhanced by adaptive histogram equalization.And then,on the basis of the previous result,the details are locally enhanced by improved wavelet transformation.Experiments show that the proposed method could stretch the scope of gray-scale values,abate noise and enhance hazy bone fissure,the results are similar to the normal image and have good visual qualities.PNSR and the number of effective gray-scale values make further proof.

关 键 词：CR图像 非线性函数 直方图均衡 小波变换 软阈值 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]