多传感器图像融合技术及其进展

The Technique and Development of Multi-sensor Image Fusion

数据融合技术是一个多级、多层面的数据处理过程,主要完成对来自多个信息源的数据进行自动检测、关联、相关、估计及组合等的处理,得出最终结论或决策的技术。多传感器数据融合是一门新兴的技术,是对数据的综合与处理的过程。该技术在军事领域和民用领域都有很大的发展和应用前景,是目前受到普遍关注并迅速发展的技术分支,本文简述了图像融合技术的原理与架构、技术特点以及像素级、特征级和决策级图像融合的分类方法,介绍了各种方法的一些具体实施方案、特点及新的进展,评述了图像融合技术进一步发展和应用的一些关键问题,并指出了多源传感器数据融合研究中存在的问题及研究方向。

Data fusion technology is a muhiple level data processing to complete the data auto examination, conjunction, correlation, estimation, as well as combination processing for the data from the multiple resources, by which the ultimate conclusion and decision could be obtained. Multi - sensor data fusion is a rising technology for the data integrating and processing. This technology has had great development and application prospect both in the military and civilian field, and is drawn great attention as well as fast booming. This paper introduced the principle and structure of the data fusion technology, the technical characteristics, as well as the three classification methods of data fusion：Pixel - lever, Feature - lever and Decision - level. It also introduced some detailed implementation plan, characteristic and progress of every method, and elaborated some key problems of the future development and application of image fusion technology, as well as indicated the exiting problems and study direction in the multi -sensor data fusion research.