Multimodal image fusion with joint sparsity model

Image fusion combines multiple images of the same scene into a single image which is suitable for human perception and practical applications. Different images of the same scene can be viewed as an ensemble of intercorrelated images. This paper proposes a novel mul-timodal image fusion scheme based on the joint sparsity model which is derived from the distributed compressed sensing. First, the source images are jointly sparsely represented as common and innovation components using an over-complete dictionary. Second, the common and innovations sparse coefficients are combined as the jointly sparse coefficients of the fused image. Finally, the fused result is reconstructed from the obtained sparse coefficients. Furthermore, the proposed method is compared with some popular image fusion methods, such as multiscale transform-based methods and simultaneous orthogonal matching pursuit-based method. The experimental results demonstrate the effectiveness of the proposed method in terms of visual effect and quantitative fusion evaluation indexes. C 1 Introduction Recently, with extraordinary advances in sensor technology, numerous imaging sensors have been developed in military and civilian applications. The images provided by different sensors of one scenario often present complementary information. An image fusion technique can integrate information from different sensors into a single image. The fused image can preserve the relevant information and reduce the uncertainty and redundancy. Compared with the image provided by the individual sensor, the fused image has several benefits: broadened the spatial and temporal resolution, improved reliability, and increased robustness.