The Infrared And Visible Image Fusion System Based On FPGA

Infrared and visible image fusion is an important branch of image fusion. By combining two images of different wavelengths, it can make us detect target information more rapidly and accurately, and improve our ability of perceiving image information. It is widely used in military and civilian applications. Therefore, it has important theoretical and practical significances for researching the infrared and visible image fusion system. Infrared and visible image fusion is the study object in this paper. Firstly the fundamental theory of image fusion is explained and fusion algorithm is designed. Then set up a fusion system based on FPGA and design the logic of FPGA. Finally, a complete solution of image fusion system is proposed.The imaging principle of infrared and visible detector is studied on the basis of understanding the overall system architecture of image fusion in this paper. Thus the infrared and visible imaging characteristics are confirmed. The image fusion pre-processing, including non-uniformity correction, image de-noising, image enhancement, image registration and so forth, is decided by the imaging principle of detectors and differences of imaging characteristic. Finally, quality assessment of the fused image is reviewed. Four methods belonging objective quality assessment are mainly introduced, which is information entropy, average gradient, edge information preservation and the structure similarity.In order to solve the contradiction of existing fusion algorithms between fusion effect and real-time implementation, this paper proposes an algorithm for infrared and visible images fusion based on feature extraction. Firstly the algorithm improves the mathematical morphological top-hat operation to extract the feature and background from the source image. Then the fusion rules are designed to fuse the feature and background images. Finally the fused image is obtained through additional operation. The experiment indicates that the proposed method can efficiently extract feature information from source image and provide abundant background information. Besides, this method can accelerate operation, so it is suitable to implement in hardware.In order to realize the real-time image fusion processing between infrared image from GWIR 03 02 XIA and visible image from KAC06040, firstly the functional requirement of fusion system is analyzed, and on this basis, the image fusion system is built based on XC6SLX150, which is a certain type of Xilinx’s FPGA. On the platform of image fusion system in this paper, complete the FPGA design by modular design method and perform a series of experiments. First of all, the validity of each module is validated by the simulation experiment. Secondly, complete the test of system hardware. Finally, design experiment of the fusion system, and then the experimental results are analyzed. The experimental results indicate that the image fusion system is established successfully. The system has satisfactory performance and can achieve real-time processing. Besides, it has the lower rate of FPGA resources occupation. In this paper, the design has kept design margin in these aspects: resource, response time.To sum up, design scheme of image fusion system in this paper is effective and feasible. In addition to finishing the design requirement, the system can be still used to the image fusion pre-processing and the fusion processing of other wavelengths. After further experiments and design optimization, the system in this paper can be applied to engineering and has certain reference significance for the study of image fusion technology.