基于卷积稀疏表示的红外与可见光图像融合

稀疏表示是以块为单位进行编码的,因此破坏了图像块间的相关性。针对上述问题,提出了基于卷积稀疏表示的红外与可见光图像融合算法。该算法采用交替方向乘子算法(ADMM)求解非下采样轮廓波变换(NSCT)域强边缘子带的卷积稀疏系数,完成特征响应系数的融合。同时,采用脉冲耦合神经网络(PCNN)模型的点火图完成NSCT域高频子带的融合。实验结果表明:该算法解决了稀疏表示的“块效应”问题,同时又兼具PCNN模型的视觉特性,可以有效地捕捉源图像的特征信息。另外,在主观视觉评价和客观质量评价方面均优于现有算法。

Infrared and visible image fusion based on convolution sparse representation

Sparse representation is encoded in blocks,which destroys the correlation between image blocks.To solve the above problem,an infrared and visible image fusion algorithm based on convolution sparse representation is proposed.The algorithm uses the alternating direction multiplier algorithm(ADMM) to solve the convolution sparse coefficients of the strong edge subbands in the non-subsampled contourlet transform(NSCT) domain and complete the fusion of the feature response coefficients.At the same time,the high-frequency subbands fusion in the NSCT domain is completed using the ignition map of a pulse coupled neural network(PCNN) model.The experimental results show that the algorithm solves the "block effect" problem of sparse representation,and simultaneously has the visual characteristics of the PCNN model,which can effectively capture the feature information of the source image.In addition,subjective visual evaluation and objective quality evaluation are better than existing algorithms.