激光与光电子学进展, 2015, 52 (10): 101004, 网络出版: 2015-10-08
S-PCNN与二维静态小波相结合的遥感图像融合研究
Remote Sensing Image Fusion Algorithm Based on S-PCNN and Two-dimensional Stationary Wavelet Transform
图像处理 遥感图像融合 简化脉冲耦合神经网络 二维静态小波 HSV 彩色空间 image processing remote image fusion simplified pulse coupled neural network two-dimensional stationary wavelet transform HSV color space
摘要
在色度、饱和度、纯度(HSV)彩色空间,结合简化脉冲耦合神经网络(S-PCNN)与二维离散静态小波(SWT)提出一种有效的遥感图像融合算法。将多色光谱转换到HSV 色彩空间,对多色光谱的V 分量与全色光谱进行二维静态小波分解,再将分解后的高频系数输入S-PCNN 模型进行融合。低频部分进行第二次小波分解并采用不同规则将其融合,对融合的小波系数进行小波逆变换得到融合的V 分量,并将多色光谱的H、S 与融合后的V 分量转换到RGB 空间。通过一组常用的遥感图像融合实验,表明本文算法的融合效果优于传统算法,且融合图像细节明细、色彩保留较好,是一种有效的遥感图像融合算法。
Abstract
In hue, saturation and value (HSV) color space, an effective remote sensing image fusion algorithm is proposed combining with simplified pulse coupled neural network (S-PCNN) and two-dimensional stationary wavelet transform (SWT). The multispectral transformed into HSV color space, the multispectral V component and the panchromatic spectrum are decomposed by two-dimensional static wavelet decomposition, and the decomposed high-frequency coefficients is put into S-PCNN model to fuse.The low-frequency coefficients are decomposed second time and fused with different rules, the fused V component is obtained through wavelet inverse transformation for fused wavelet coefficient, the multispectral H, S components and fused V component are transformed into RGB space. Through a group of common remote sensing images experiment, the results show that the fusion effects of proposed algorithm is better than the traditional algorithms, and the fused image contains lots of detail, color. It is an effective remote sensing image fusion algorithm.
金鑫, 聂仁灿, 周冬明, 余介夫, 贺康健. S-PCNN与二维静态小波相结合的遥感图像融合研究[J]. 激光与光电子学进展, 2015, 52(10): 101004. Jin Xin, Nie Rencan, Zhou Dongming, Yu Jiefu, He kangjian. Remote Sensing Image Fusion Algorithm Based on S-PCNN and Two-dimensional Stationary Wavelet Transform[J]. Laser & Optoelectronics Progress, 2015, 52(10): 101004.