红外技术, 2019, 41 (1): 44, 网络出版: 2019-03-23  

面向图像差异特征融合的基于弗里德曼检验的小波基分类研究

Wavelet Bases Classification Research Based on Friedman Test for Image with Difference Features Fusion
作者单位
中北大学信息与通信工程学院, 山西太原 030051
摘要
对小波基分类是图像融合中依据不同融合需求选择小波基的基础, 可以提高图像融合的智能化水平。针对现有的小波基分类方法仅根据小波自身特性进行分类, 没有从统计角度有效建立小波基和图像差异特征之间的联系, 本文提出了面向图像差异特征融合的基于弗里德曼检验的小波基分类方法。首先, 选择典型的差异特征和小波基用于分类研究; 其次, 选择针对差异特征的评价指标, 以评价指标结果作为标记量并进行分类实验的区组设计; 然后, 采用弗里德曼检验对不同区组数据进行处理及执行相应的后续检验和分类步骤, 形成面向图像差异特征的小波基类集; 最后, 设计对比实验对分类方法的有效性进行验证和分析。试验结果表明, 该分类方法能有效把对图像差异特征融合效果相近的小波基归为一类, 能根据融合需求选择较好的小波基。
Abstract
The classification of wavelet bases is in accordance with different fusion requirements in the image fusion process, which can improve the intelligence level of image fusion. In the existing wavelet base classification methods, based only on the wavelet’s inherent characteristics, there is no effective connection between the wavelet bases and the image difference features from the statistical perspective. This paper proposes a wavelet base classification method based on Friedman test for the image feature fusion combining different features. Initially, typical difference features and wavelet bases for classification research are selected, and then evaluation indices to evaluate difference features are selected. The evaluation indices results are utilized as mark measures to conduct block design of the classification experiment. Subsequently, the Friedman test to process different data blocks is utilized and the corresponding test procedures are implemented. Furthermore, classification procedures are performed to construct a set of wavelet bases that can fuse image difference features satisfactorily. Finally, a comparative experiment is designed to verify and analyze the effectiveness of the classification method. The experimental results demonstrate that the classification method can effectively classify the wavelet bases with similar fusion performance on image difference features into subsets and can choose a better wavelet base according to fusion requirements.

王向东, 杨风暴, 焦玉茜, 吉琳娜, 吕红亮. 面向图像差异特征融合的基于弗里德曼检验的小波基分类研究[J]. 红外技术, 2019, 41(1): 44. WANG Xiangdong, YANG Fengbao, JIAO Yuqian, JI Linna, LYU Hongliang. Wavelet Bases Classification Research Based on Friedman Test for Image with Difference Features Fusion[J]. Infrared Technology, 2019, 41(1): 44.

关于本站 Cookie 的使用提示

中国光学期刊网使用基于 cookie 的技术来更好地为您提供各项服务,点击此处了解我们的隐私策略。 如您需继续使用本网站,请您授权我们使用本地 cookie 来保存部分信息。
全站搜索
您最值得信赖的光电行业旗舰网络服务平台!