象素级多传感器图像融合技术

Pixel level Multisensor Image Fusion Technique

多传感器图像融合是图像信息的综合与分析过程，它可有效地萃取多通道图像信息，构成统一图像，提供显示中心显示或决策机构判别。各种图像传感器的性能极大地依赖于使用环境，至今还没有一种图像传感器在多种不同的使用条件下，其各项性能（作用距离、测量精度、分辨率、抗干扰、灵敏度等）都高于其它类型的图像传感器，融合多传感器图像能够提高图像探测系统的探测和识别能力。该文论述了象素级多传感器图像融合的概念、原理、处理结构。指出象素级多传感器图像融合中存在的主要问题及解决方法。利用微光ＣＣＤ摄像机和滤光片的适当组合，配以计算机的处理，在夜间对野外静态景物完成了多传感器（或多光谱）图像融合试验，结果表明。

Multisensor image fusion is a process of synthesising and processing of multisensor image data. The composite image is got by fusing many channels of images, and displayed in display center and judged by decision authorities. The performance of all kinds of image sensisors depends much on working conditions and there is not any of image sensors whose performance (working distance, measurement precision, resolution, anti interfere sensitivity, etc.) is bettter than other kinds of image sensors under different working conditions. Multisensor image fusion can improve the capability of image detection system for detecting and recognizing targets. Concept, principle, processing construction for pixel level multisensor image fusion are described, and major problems and possible ways of solution are also presented. The proper combination of low light level(LLL) camera with optical filters is used to implement multisensor(or multispectral) image fusion experiment at night for outdoor static scene by the aid of computer. The results of experiment indicate that multisensor image fusion has improved the capability for recognizing targets.