评价彩色图像自动聚焦清晰度的互补色小波测度  

A Complementary Color Wavelet-based Measure on Color Image Sharpness Assessment for Autofocus

作　　者：周汶 李旦 张建秋 ZHOU Wen;LI Dan;ZHANG Jian-Qiu(Department of Electronic Engineering and the Research Center of Smart Networks and Systems,School of Information Science and Technology,Fudan University,Shanghai 200433)

机构地区：[1]复旦大学信息科学与工程学院智慧网络与系统研究中心和电子工程系,上海200433

出　　处：《自动化学报》2020年第8期1615-1627,共13页Acta Automatica Sinica

基　　金：国家自然科学基金(11827808,11974082);上海市科技创新行动计划社会发展科技领域项目(19DZ1205805);上海航天科技创新基金;珠海复旦创新研究院项目资助。

摘　　要：针对彩色图像的自动聚焦,本文提出了一种新的清晰度评价测度.该测度借助于互补色小波变换,在互补色小波域,利用融合互补色算子的层级最大能量和层级统计分布扩散度的乘积来描述本文的清晰度.分析表明:融合互补色算子,可提取待评价彩色图像在颜色、亮度、方向、尺度和各通道分量间的相互信息等方面的相关特征.这样,其层级的最大能量就反映了这些被提取相关特征中最显著特征的清晰程度,而其统计分布的扩散度,就衡量了其清晰度特征分布的离散程度.那么利用它们来共同表征清晰度的测度,就使得本文所提的测度能随图像清晰程度的增加而增加.在LIVE/IVC数据库上与多种经典方法的对比结果表明:本文提出的测度具有最高的聚焦精度0.0373/0.0246、分辨率1.6132/0.4771和最好的无偏稳定性.In this paper,a new measure on color image sharpness assessment for autofocus is proposed.By means of the complementary color wavelet transform(CCWT),the scale maximum energy and statistical distribution variabilities of the fusion complementary color operators in CCWT domain are combined into an image sharpness assessment measure by their product.The analyses show that the complementary color operators can extract the characteristics of the color,brightness,scales,directions and mutual information among the channels in a color image.In this way,the sharpness of the most significant feature in these extracted characteritics is revealed by the scale maximum energy of the complementary color operators.The dispersion degree of the features related to the sharpness of an image is disclosed by the distribution variabilities of the operators.As a result,the proposed measure in this paper can increase with the increase of the image sharpness.The simulation results on the LIVE/IVC database show that the proposed measure is of the highest focusing accuracy 0.0373/0.0246 and resolution 1.6132/0.4771,and the best unbiased stability comparison with the classical methods reported in literature.

关 键 词：清晰度评价 互补色小波 彩色图像 最大能量 扩散度 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]