APP下载

改进的PCNN模型在多光谱与全色图像融合中的应用研究

2014-03-05段先华操佳佳刘佳

现代电子技术 2014年3期

段先华+操佳佳+刘佳

摘 要: 介绍了PCNN模型原理,提出了基于双通道自适应的PCNN多光谱与全色图像融合算法。该算法首先将RGB空间的多光谱图像转换为HSV彩色空间,然后将HSV彩色空间中的非彩色通道(V通道)的灰度像素值和全色图像的像素灰度值分别作为PCNN?1及PCNN?2的神经元输入,利用方向性信息作为自适应链接强度系数,对非彩色通道图像和全色图像进行自适应分解,再将点火时间序列送入判决因子得到新的非彩色通道图像,最后将原多光谱图像的H通道分量、S通道分量及新的V通道分量经HSV空间逆变换获得最终的融合图像。实验结果表明,该算法不仅解决了链接强度系数自动设置的问题,而且充分考虑到图像边缘和方向特征的影响,无论在主观视觉效果,还是客观评价标准上均优于IHS、PCA、小波融合等其他图像融合算法,同时降低了计算复杂度。

关键词: 遥感图像; 图像融合; PCNN模型; HSV彩色空间

中图分类号: TN911.73?34 文献标识码: A 文章编号: 1004?373X(2014)03?0055?06

Application research of modified PCNN model in multispectral

and panchromatic images fusion

DUAN Xian?hua, CAO Jia?jia, LIU Jia

(College of Computer Science and Engineering, Jiangsu University of Science and Technology, Zhenjiang 212003, China)

Abstract: The principle of PCNN model is introduced, and an image fusion algorithm of multispectral and panchromatic based on adaptive dual?channel PCNN is proposed. Firstly the multispectral image of RGB space is converted to HSV color space, and then the gray level of achromatic channel (channel V) in HSV color space and gray level of panchromatic images are entered as PCNN?1 and PCNN?2 neurons respectively, using orientation information as an adaptive link strength factor to achieve adaptive decomposition for achromatic channel and panchromatic images. The duration of ignition sequence is sent to judgment factor to get new of achromatic channel image. Finally the final fused image is obtained through HSV space inverse transformation of H channel component, S Channel component and new V?channel component of original multispectral image. Experimental results show that the algorithm not only solved the problem of automatically setting linking strength factor, but also took full account of the effect of image edge and directional characteristics. No matter the subjective visual effect, or objective evaluation standards they are all better than other image fusion algorithms such as IHS, PCA, wavelet transform image fusion, the computational complexity is reduced in the meantime.

Keywords: remote sensing image; image fusion; PCNN model; HSV color space

0 引 言

多光谱图像具有较高的光谱分辨率和较低的空间分辨率,全色图像的空间分辨率较高,但具有较低的光谱分辨率。在遥感应用领域,有时会要求图像的空间分辨率和光谱分辨率都比较高,但由于技术条件的限制,很难提供这样的数据。因此,为了获得较高的空间分辨率同时保留多光谱图像的光谱特性,将多光谱图像和全色图像进行融合成为了图像融合领域的研究热点。自 20 世纪研究者们开始对遥感图像融合技术关注以来,针对多光谱与全色图像融合提出了大量算法,大致可分为两大类[1]:基于彩色空间分量替换的融合方法和基于多分辨率分析的融合方法。……

登录APP查看全文