基于感受野模型的图像融合算法研究  

An Image Fusion Algorithm Based on Receptive Field Model

作　　者：赵巍[1] 黄晶晶[1] 田斌[2] 

机构地区：[1]北京航空航天大学203教研室,北京100083 [2]第二炮兵装备研究院,北京100085

出　　处：《电子学报》2008年第9期1665-1669,共5页Acta Electronica Sinica

基　　金：国家自然科学基金资助项目(No.60502019)

摘　　要：现实世界中最优的图像融合系统莫过于生物视觉系统.如果能够了解生物视觉的基本原理,建立其数学模型,并能应用于多传感器图像融合实践中,将会对多传感器图像融合技术的发展有很大促进作用.本文以人类视觉的生物机理和感受野的数学模型为基础,提出了一种新颖的基于人眼视觉感受野模型的图像融合算法.这种算法非常适合融合两幅灰度差异较大的异质传感器图像,融合图像在保留两幅原图像中有用信息的同时,还对图像间的差异进行了增强,有助于提高观察者识别目标的速度和准确性.另外,该算法还有一定的去噪功能,计算量很小,可满足实时性要求.The best image fusion system in real world is human vision.If we can know the biological machine and build its mathematic model, it will be very useful in application. In this paper, we describe the biological knowledge of human eyes, such as receptive field. And a new image fusion algorithm based on human vision model is presented. This method is suitable to fuse two images with different gray levels fi＇om different sensors. While all usefui infomaation in original images can be observed in fused images, the differences of two images are enhanced. The enhancement helps the observer to recognize the objects correctly and quickly. Additionally, this algorithm reduces the noise in original images and has low calculation cost. It can satisfies the real lime requirement.

关 键 词：图像融合 感受野 生物视觉 

分 类 号：TP91[自动化与计算机技术]