模糊数学和投影至闭凸集法结合提高医学图片谱估计质量

Improving of Hedical Image Spectrum Estimation by Combing Fuzzy Sets and POCS Methods

传统的功率谱估计得到的估值可视为是真实功率谱在隐含窗和噪声作用下退化的结果,可采用信号解卷积的方法恢复信号的真实功率谱。本文采用模糊数学和投影至闭凸集(POCS)法相结合实现解卷积的选代算法,对约束条件和隶属函数的选择均作了考虑。采用在模糊集的截集上的选代求解,提高了谱估计质量。该方法用于图片的功率均衡滤波,获得较好的效果。

The estimation of power spsctrum by classical method ca be regarded as the result of real power spectrum degr aded by noise and implicit window It can be restored by the method on decovoluting signal. This paper gives an algarithm of iteration and deconvolution by combining the Fuzzy Sets and the POCS(Project on Convex sets) method. The restraint and the choice of attribute function has been considered. The quality of spectrum estimation is improved by iterative aigarithm on Fuzzy Sets. It was used for the power spectrum balance filtering of image and a satisfy result was obtained.