提出了一种新的图像质量快速盲检测方法,以锐利度作为图像质量准则,而锐利度可以归结为图像的模糊特性,并由边缘信息估计得到。为了提高系统效率,采用边缘对比度的估计方法,并以边缘对比度作为边缘锐利度的判据准则。分析了邻近边缘的相互影响,因而只需从有限的边缘点来提取出合理的线扩展函数(LSF),进而得到模糊参数。在LSF计算中,采用一种新的插值计算方法。实验结果表明:当图像不是特别模糊时,测量结果是准确的。
A new metric for image quality assessment is presented in terms of image sharpness while the sharpness is considered as characteristics of blurring.The blur parameter is estimated using edge information.To improve the system efficiency,a criterion for edge sharpness is employed and only the sharpest edge is selected for the extraction of line spread function(LSF).The effect of nearby edges on edge feature and LSF is analyzed and a space constrain is put forward to select appropriate LSFs.A new method for pixel interpolation is presented in LSF computation.The experimental results demonstrate that the proposed method is accurate when the blur is not serious.