基于模板匹配法的散焦图像模糊度估计.pdfVIP

ComputerEngineeringandApplications计算机工程与应用 基于模板匹配法的散焦图像模糊度估计 朱宏亮，张 荣，尹 东 ZHU Hong·liang，ZHANG Rong，YIN Dong 中国科学技术大学 电子工程与信息科学系 信息处理中心重点实验室，合肥 230027 Key Laboratory ofInformation Processing Center，DepartmentofElectronic Engineering Information Science．University of Science Technology ofChina，Hefei230027，China E—mail：zhuhl@mail．vstc．edu．cn ZHU Hong—liang，ZHANG Rong，YIN Dong．Template matching based blur estimation method for out-of-focus images． ComputerEngineeringandApplications，2010。46(36)：184-187． Abstract：Thispaperproposesanew methodtoestimatetheparameterofPointSpreadFunction(PSF)forout—of-focuslm— age．Atfirst，the out—of-focusblurring ismodeled as Gaussian blurring．Then，analyze the differentialfeatures ofseveraltypes ofimageedgesand extractGaussianLineSpreadFunction(LSF)from imageedgesusingthesedifferentialfeatures．Since matching filtering hasthe advantage ofmaximizing the outputsignal—to—noise ratio，based on thisphilosophy this paperap— pliestemplatematchingmethodto theevaluation ofLSF standarddeviation(STDEV)．The STDEV ofGaussianPSF can be easily calculated from the STDEV of LSF by the symmetric property of PSF．The resultshows thatthe proposed method canaccuratelyevaluatetheSTDEV ofout—of-focusimagePSF．Finally，thispapertakestheSTDEV ofPSFasanobjective metricofimageblurriness．Thetestingresultshowsthatthisobjectivemetriccorrelateswellwiththejudgmentofhumanob— servers．Therefore，the STDEV ofPSF isan excellentquality metric ofrout—of-focusblurimages． Keywords：out—of-focusblur；pointspread function；linespread function；image blurriness；templatematching 摘 要：提出一种新的散焦图像模糊度估计方法。由图像中的阶跃型边缘提取线扩展函数。根据匹配滤波可实现最优信噪比信 号检测的思想，引入模板匹配的方法计算线扩展函数的标准差。由点扩展函数的圆周对称性从线扩展函数的标准差得到高斯型 点扩展函数的标准差。实验结果表明，该方法能够准确地估计散焦模糊图像点扩展函数的标准差。将点扩展 函数的标准差作为 一 种评价图像模糊度的测度。实验表明该测度符合人眼的视觉特性，可以很好地判定散焦模糊图像的模糊度。 关键词：散焦模糊；点扩展函数；线扩展函数 ；图像模糊度；模板匹配 DOI：10．3778~．issn．1002—8331．2010．36．05l 文章编号：1002．8331(2010)36—0l84．04 文献标识码：A 中图分类号：TP39 1 引言