基于小波变换的实时图像融合技术的实现  

The implement of real-time image fusion based on wavelet-analysis

作　　者：李剑[1,2] 张覃平[1] 

机构地区：[1]中科院光电技术研究所,成都610209 [2]中国科学院研究生院,北京100039

出　　处：《仪器仪表用户》2010年第3期56-58,共3页Instrumentation

摘　　要：提出了一种针对可见光和红外图像实时融合的系统。该系统由成像和图像处理两部分组成。其中,成像部分使用同一孔径以减少视差效应,并采用分光器以使同一光路光分别在红外和可见光传感器上成像。在图像处理部分,首先对两路图像进行配准、同步处理,然后在FPGA内部实现了一种基于小波分析的图像融合方法,实现了对两路视频图像实时进行融合。和Matlab中双精度运算结果相比,该系统的输出结果可达约43dB的信噪比,满足实际的工程应用。A real-time image fusion system is presented for the fusion of visible and infrared images. This system consists of two parts, the imaging configuration and the image processing unit. In the imaging configuration, a common aperture is used to avoid, the parallax effects, and the beam-splitter is employed to allow the infrared and visible sensors to share a common optical path. In the image processing unit, images are first registered and frame rate is synchronized, and then a fusion method based on wavelet-analysis decomposition is implemented on FPGA. The results outputted from the system achieve the signal-to-noise ratio （SNR） about 43dB. This demonstrates that the system can be applied to the practical applications.

关 键 词：红外图像 图像融合 实时系统 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]