三种频闪自动检测算法的实现及对比分析研究  

Implementation and comparison of three automatic flicker detect methods

作　　者：陈孟儒[1] 张宇弘[1] 

机构地区：[1]浙江大学超大规模集成电路设计研究所,浙江杭州310027

出　　处：《机电工程》2008年第11期16-19,共4页Journal of Mechanical & Electrical Engineering

摘　　要：为解决交流光源给手机、笔记本电脑等移动电子产品中的基于电子滚动快门的互补型金属氧化物半导体(CMOS)图像传感器带来的频闪问题,提出了频闪自动检测的改进方案及新方案。阐述了三种频闪自动检测的算法(分别是:频率分量提取方式,幅度调制解调方式,Harr小波方式)及其超大规模集成电路(VLSI)实现,并对三种方案在性能及实现代价上进行了比较。分析结果显示,Harr小波方式在性能及代价上相对其他两种方法较好。Aiming at solving the flicker problem of eomplementory metal-oxide-semiconductor（CMOS） image sensor based on e- lectronic rolling shutter which is widely used in mobile electronic systems such as cell phones and laptops caused by AC light, an improvement and new algorithms were proposed to detect flicker automatically. The three methods and their very large scale inte- gration（VLSI） implementation were presented, including frequency distill method, amplitude modulation demodulate method and Hart wavelet method. The comparison shows that the third one is the best in both performance and cost.

关 键 词：频闪自动检测 电子滚动快门 互补型金属氧化物半导体 图像传感器 超大规模集成电路 

分 类 号：TP331.1[自动化与计算机技术—计算机系统结构]