远程采集随即视频图像清晰化方法仿真  

Simulation on Sharpening Remote Acquisition Video Image

作　　者：李艳平[1] 杨格兰[2] 

机构地区：[1]菏泽学院计算机与信息工程系,山东菏泽274015 [2]中国科学技术大学自动化系,安徽合肥230027

出　　处：《计算机仿真》2013年第9期230-233,285,共5页Computer Simulation

摘　　要：远程采集的图像中,存在较强的随机性因素,图像中包含的噪声因素不可控,视频图像受到亮度不定、意外抖动、图像冻结等不可以抗拒的因素影响,造成像素缺失,清晰度下降,传统的以评价函数为基础的去噪方法仅仅针对像素这一单模态的缺损进行修补,没有考虑色差、偏色等多因素的影响,可靠性和准确性不强。提出一种多模态的视频图像清晰化方法。改进方法融合了视频图像清晰度、偏色、噪声等多个特性检测属性。运用一种分块平方梯度聚焦评价函数进行视频图像的清晰度检测,采用一种等效圆的偏色检测方法,增加对随机视频图像的偏色检测,融入了一种改进的Minmax噪声检测算法辅助清晰化的效果。实验结果表明,改进的方法具有较好的可靠性,提高视频图像清晰度。This paper presented a video image clarifications method based on multi-mode,which blends some algorithms of video clarity detection,color cast detection,and image noise detection.The evaluation function based on blocking square gradient focusing was used to detect clarity of a video,which has higher credibility and accuracy,compared with some former common evaluation methods.Detecting the color cast based on equivalent circle was supposed,which carried out the function of video color cast detection.Improved Min-max algorithm was utilized to detect image noises.Experimental results demonstrate that improved method has better reliability and the video images are more clarity.

关 键 词：多模态 视频诊断 清晰度 偏色 图像噪声 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]