安检中X光断层扫描加速迭代算法研究

随着安检要求的日益提高,计算机断层扫描的安检方式正成为X光安检中越来越重要的部分.为了保证行李扫描的检测效率,加快断层扫描图像的迭代计算速度,提出了一种采用小波变换域取代图像空间来测算图像小波系数的估值方法.该方法以交替最小化算法为框架,在逼近系数时,创建一个替代函数,随着迭代次数的增加,基于同一标准的小波树得以更新扩展,每层的细节系数均更新,小波树也依此扩展,以此达到加速迭代的效果.实验结果表明,该方法比原交替最小化算法最大能减少66.7％的迭代运算次数.

Research on Accelerating Iterative Algorithm of Security Screening by X-Ray Tomography

With the requirement of security increasing day by day,the computed tomography (CT) scan mode of X-ray inspection is becoming more and more important part of the security.In order to ensure the detection efficiency of baggage scanning and speed up the iterative calculation of tomography image,a method using the wavelet transform domain instead of the image space to measure the image wavelet coefficients of the valuation is proposed.The method takes the alternating minimization algorithm as framework and while be founded in approximation coefficients,an alternative function is created.With the increase of the number of iterations,the wavelet tree based on the same standard can be updated and expanded,in which detail coefficients of each layer are updated and wavelet tree also depends on the extension to accelerate the effect of the iteration.Experimental results show that the presented algorithm can reduce at most 66.7% of the iterative operation times compared with the alternating minimization algorithm.