【摘要】 目标识别是计算机视觉领域重要的研究课题。图像融合由于能对信息进行多级别、多层次以及多方面的处理与综合,因此可以获得更为精确和更为可靠的信息。近年来,基于多特征融合的目标识别方法由于无需精密的图像配准,计算简便的优点得到研究者的重视。本文首先简要介绍了特征融合的基本原理和基本框架;然后梳理了特征融合的几类重要方法;最后,经过对前人工作的总结和分析,结合我们的工作,指出了现有算法的不足之处,并对今后研究中的发展方向和挑战提出了预测性的看法和建议。更多还原

【Abstract】 The object recognition is an important research topic in the field of computer vision.However,image fusion is the multi-level and comprehensive method in information processing and synthesizing,so we can get more accurate and more reliable information.In recent years,target recognition based on multi-feature fusion attracts many researchers’attention because of its low image registration precision and the simple computation.We firstly introduce the basic principles and framework of feature fusion;then review the important feature-level fusion methods.Finally,after the summary and analysis of previous work and our current works,we point out the inadequacies of the existing algorithms,predict the future research direction and challenges,and give some views and suggestions.更多还原