面向人眼探测识别的视频优化方法研究  

Research of Video Optimization for Detection and Recognition by Human Eyes

作　　者：吴锦[1,2] 范晓鹏[1,2] 朱枫[2] 

机构地区：[1]中国科学院研究生院,北京100049 [2]中国科学院沈阳自动化研究所,辽宁沈阳110016

出　　处：《机械设计与制造》2013年第1期101-103,共3页Machinery Design & Manufacture

摘　　要：在人眼探测识别过程中,观察者总是希望从图像中获取尽量多的信息以用于探测识别。因此,国内外学者提出了很多图像增强和优化的算法来处理图像,使其更适宜人眼观察。虽然这些算法处理单幅图像时可以获得很好的效果,但在处理视频或序列图像时会遇到图像闪烁的问题。这主要是由于算法在处理前后几帧时,在同一区域的像素灰度值出现了很大的跳变。在之前图像优化算法研究的基础上,首先分析其用于视频优化时出现闪烁的原因,然后提出一种在保证图像信息能最大限度的为人眼感知的基础上,抑制视频闪烁的方法。In the process of detection and recognition by human eyes,the observers always hope to get information from the images as much as possible.Therefore,researchers at home and abroad proposed a lot of image enhancement and optimization methods to process the image,so that it is more suitable for observation.These methods,though,are able to get good result when used to process a single image,but as processing video or a series of images,they may encounter the problem of image flickering.The main reason lies in that pixels in the same region of continuous frames get a big step in gray level after the methods processed them.The research which based on previous one on image optimization method,analyzes the reason of flicker in video optimization firstly,then proposes a method to suppress the flicker in video on the premise that human can get the maximum information from it.

关 键 词：视频优化 闪烁 探测识别 

分 类 号：TH16[机械工程—机械制造及自动化]