Ir, Light/visible Light Image Fusion Algorithms

Multi-source image fusion technology make full use of the correlation of different sensors on the space-time and the complementary information, combine the feature of multi-spectral images in the same scene together, then give a more comprehensive and clear description of the scene. It has a good prospect in military operations, aerospace, remote sensing, medical diagnostics etc. In this paper, we studied the algorithm of infrared image and LLL image/visible image fusion. The specific contents are as follows:Analyzed the characteristics of the targets in the different spectral environment. The target reflection/radiation contrast model, atmospheric spectral transmission model, the detector spectral matching model were constructed, eventually proposed a new weight calculation model of LLL and infrared image fusion.For unequal spectral information of infrared image and LLL/visible image, we proposed an image fusion algorithm region-based. First, the source images are decomposed, and the residual images are fused by weighted average method. Fuzzy C-Means clustering algorithm is used to Segment. Then, Regional mapping is carried out in each sub-layer. Through the corresponding fusion criterions, fusion is completed. The region-based color transfer algorithm is introduced. Simulation experiments prove the effectiveness of the algorithm.Morphological top-hat operation, image variance, Chebyshev inequality are used to complete the extraction of the target mask information. Then taking the intersection of the infrared image with the target mask information, As the corresponding positions in the fused image. Select the complement of the target mask information with LLL image/visible image, as the non-target location. The model of image fusion algorithm based on the visual cognition is built. Simulation experiments prove that the proposed algorithm can highlight the target information in the human visual stimulation effect without loss of the details.Using a top-down design methodology, we design the software of the multi-source grayscale/color image fusion and evaluation under the Matlab GUI environment. Collected the typical representative algorithm from the beginning of image fusion to now. it is the summary of the study and work during my master. The software have some realities value to follow-up study and improvement for the image fusion algorithm.