基于小波变换的图像融合算法研究

2018-04-04 09:10:16刘娜田大为
现代电子技术 2018年7期

刘娜 田大为

摘 要: 图像融合是改善图像质量的一个重要途径,传统算法难以正确地对图像进行有效融合。为了提高图像融合的质量,提出一种基于小波变换的图像融合算法。首先对2幅原始图像进行小波变换,提取它们的小波系数,然后采用不同的规则对不同层次的小波系数进行融合,并采用小波变换对融合的系数进行融合,最后采用不同类型对图像融合结果进行测试和分析。结果表明,小波变换融合后的图像更加自然、清晰,提高了图像的信噪比,并且图像融合速度明显加快,获得了比对比算法更加理想的图像融合效果。

关键词: 图像质量; 小波变换; 神经网络; 加权融合; 信噪比; 图像融合

中图分类号: TN911.73?34; TP391 文献标识码: A 文章编号: 1004?373X(2018)07?0071?04

Research on image fusion algorithm based on wavelet transform

LIU Na1, TIAN Dawei2

(1. Center of Health Administration and Development Studies, Hubei University of Medicine, Shiyan 442000, China;

2. Dongfeng Affiliated Hospital of Hubei University of Medicine, Shiyan 442000, China)

Abstract: The image fusion is an important way to improve the image quality, but the traditional algorithm is difficult to fuse the image effectively. In order to improve the quality of image fusion, an image fusion algorithm based on wavelet transform is proposed. Wavelet transform is performed for two original images to extract their wavelet coefficients. The different rules are used to fuse the wavelet coefficients at different levels, and then the wavelet transform is used to fuse the fused coefficients. The images with different types are adopted to test and analyze the image fusion results. The results show that the image after wavelet transform fusion is more natural and clearer, the proposed algorithm can improve the signal?to?noise ratio of the image, quicken the speed of image fusion, and can obtain more satisfactory image fusion effect than the comparing algorithms.

Keywords: image quality; wavelet transform; neural network; weighted fusion; signal?to?noise ratio; image fusion

0 引 言

当图像采集环境不理想时,采集单一的一幅图像无法准确地描述目标信息,需要通过对目标的多幅图像进行融合,以更加全面、准确地描述目标信息。图像融合成为图像处理的一个重要方面,在遥感、气象等领域具有十分重要的实际应用价值。为此,图像融合算法的研究引起了人们的高度关注[1?3]。

针对图像融合问题,国内外专家进行了深入的分析。通常情况下,传感器对同一个目标的图像信息进行采集,并进行去噪、归一化等预处理操作,最后通过图像融合算法实现信息互补,得到一幅比原图像质量更高的图像[4?5]。最初采用单一传感器进行信息采集,但是单一传感器的缺陷十分明显,如分辨率、光谱等方面的不足。……

登录APP查看全文