Research On Application Of Shearlet Transform In Image Fusion

Image fusion is a process of fusing the important information of multiple images from different sensors to an image, which can overcome limitation and difference of the image from single sensor, obtain a more comprehensive and accurate description of the scene, improve image clarity and understandability and be beneficial to further image analysis and processing. Image fusion technology as an integrated technology, covers sensor technology, signal processing technology, computer technology and artificial intelligence and other disciplines. In recent years, it is widely used in medicine, military, remote sensing, meteorological forecast and other fields. Therefore, Shearlet transform and non-subsampled Shearlet transform are applied to image fusion in order to work on new image fusion algorithms in the dissertation.First of all, the multi-focus image fusion algorithm based on Shearlet transform is studied in the dissertation. According to the imaging characteristics of multi-focus images and the decomposition coefficient characteristics of Shearlet transform, an image fusion algorithm based on cycle spinning Shearlet transform is proposed. The fusion rule of weighting by regional SML (Sum-modified-Laplacian) value is applied to fuse the low frequency coefficients of Shearlet domain and the fusion rules of choosing the larger regional coefficient SML value is applied to fuse the high frequency coefficients of Shearlet domain. Experimental results show that the fused images obtained by the proposed algorithm have advantages of high clarity and rich detail texture information.Secondly, the infrared image and visible image fusion algorithm based on non-subsampled Shearlet transform is studied in the dissertation. According to the characteristics of infrared image and visible image and human visual properties, a kind of infrared and visual image fusion algorithm based on object extraction and adaptive pulse coupled neural network via non-subsampled Shearlet transform is proposed, which fuses the object region and background region respectively by different fusion rules. Experimental results show that the target information of infrared image and the background information of visible image of the algorithm can be well integrated in the proposed algorithm.Finally, the medical color image fusion algorithm based on non-subsampled Shearlet transform is studied in the dissertation. According to the imaging characteristics of medical image, combining with the HSI(Hue-Saturation-Intensity) color model and the, a medical color image algorithm based on non-subsampled Shearlet transform and Gaussian mixture model is proposed. Experimental results show that visual effect and objective evaluation index of the fused image by the proposed algorithm have been improved.