图像融合中，多数边缘保持滤波器在优化过程中会损坏细节和纹理信息，并且噪声也会严重影响融合结果，使得融合结果之间出现边界模糊和细节丢失问题。提出了一种基于RPCA(Robus principal compo-nent association)算法的红外光和可见光图像融合方法，可有效提高图象清晰度和视觉信息的保真度。首先，利用鲁棒主成分分析(RPCA)分解源图像为低秩部分和稀疏部分，并运用相对全变分和平均能量法对两者进行处理，最后通过NSCT逆变换获得融合图像。实验结果表明，与其他方法相比，该方法所得融合图像的平均梯度、空间频率、边缘强度、互信息量均有提升，提升量级分别为10。6％到72。6％、15％到60。2％、9。7％到69。6％，22。7％到 229。7％。
Infrared and visible image fusion based on improved RPCA algorithm
In image fusion,most edge-preserving filters corrupt structure and texture information during the opti-mization process,and the noise will also seriously affect the fusion result,which may cause problems such as loss of details and textures in the fusion results.An infrared and visible image fusion method based on RPCA(Robust princi-pal component analysis)algorithm is proposed in this paper,which can effectively improve figure definition and visual information fidelity.Firstly,infrared and visible light images were decomposed into low rank and sparse images through Robust principal component analysis.Then,relative total variation(RTV)and average energy method were adopted to process low rank and sparse images.Finally,the final fusion image was obtained by inverse NSCT transformation.The experimental results show that,compared with the other methods,the fusion image generated by the method proposed in this paper has certain improvements in the average gradient,spatial frequency,edge intensity and mutual informa-tion,with an increase of 10.6％to 72.6％,15％to 60.2％,9.7％to 69.6％,and 22.7％to 229.7％,respectively.
infrared and visible image fusionrobust principal component analysisrelative total variationaver-age energy method
NETL
NSTL
万方数据
