Computer Engineering and Applications ›› 2012, Vol. 48 ›› Issue (5): 183-185.

• 图形、图像、模式识别 • Previous Articles     Next Articles

Multi-spectral image fusion using quincunx-sampled directional filter bank

LI Yan, LIU Bin   

  1. School of Mathematics and Computer Science, Hubei University, Wuhan 430062, China
  • Received:1900-01-01 Revised:1900-01-01 Online:2012-02-11 Published:2012-02-11

采用梅花采样方向滤波器组的多光谱图像融合

李 燕,刘 斌   

  1. 湖北大学 数学与计算机科学学院,武汉 430062

Abstract: Using a characteristic that quincunx-sampled directional filter can provide rich direction information, a multi-spectral image fusion method is proposed based on the directional filter bank. An original multi-band image is converted from RGB mode to IHS mode through IHS transform. Quincunx-sampled directional filter transform is performed on the component I and the panchromatic image, corresponding low frequency sub-image and high frequency different directional sub-image are gained, and they are fused according to some fusion rules. The new I component can be obtained. A new RGB image is gotten through an inverse IHS transform. The experimental results show that the method has a better visual effect in image edge details, and has better fusion effect than IHS fusion method and the fusion method based on quincunx-sampled discrete wavelet transform.

Key words: quincunx-sampled directional filter bank, multi-spectral image fusion, quincunx-sampled discrete wavelet transform

摘要: 利用梅花采样方向滤波器能够提供丰富的方向信息的特点,提出了一种基于此方向滤波器组的多光谱图像融合方法。该方法通过IHS变换,将多波段影像从RGB空间转换到IHS空间,对I分量及全色影像进行梅花采样方向滤波器变换,得到相应的低频子图像和高频不同方向的子图像,将它们按一定的融合规则得到新的I分量,通过IHS逆变换得到RGB融合图像。实验结果表明,该方法在图像边缘等细节上有较好的视觉效果,取得了优于IHS方法和基于梅花形采样离散小波变换方法的融合效果。

关键词: 梅花采样方向滤波器组, 多光谱图像融合, 梅花形采样离散小波变换