FPGA Implementation Of Image Fusion Algorithm Based On Nonsubsampled Contourlet Transform

Image fusion is the synthesis and processing of multiple images acquired by different types of sensors from the same scene to obtain a composite image containing more information.Image fusion technology has attracted worldwide attention as it emerges,and is widely used in military,medical and aerospace fields.At present,the research of image fusion mainly focuses on algorithm research and software implementation,and lacks real-time hardware implementation research of fusion algorithm.The fusion algorithms implemented by the existing hardware fusion system are relatively simple,and it is difficult to achieve a better fusion effect.In this paper,FPGA is used as the core device,and the characteristics of image fusion algorithm based on non-subsampled contourlet transform(NSCT)are studied and the hardware simulation of the whole algorithm is realized.The specific research contents are as follows:This paper studies the FPGA implementation of visible light and infrared image fusion algorithms based on NSCT.NSCT can realize multi-scale and multi-directional decomposition of images.It is an excellent image processing algorithm and has been widely used in the field of image fusion.However,due to the complexity of its own structure,the implementation of the NSCT algorithm based on FPGA has not yet appeared.Based on the detailed description of the NSCT principle,this paper analyzes the influence of various factors(filter types,etc.)on the NSCT algorithm.At the same time,in order to balance the image fusion effect and the real-time data stream processing characteristics of FPGA,this paper selects the low-frequency coefficient fusion strategy based on the histogram saliency feature and the high-frequency coefficient fusion strategy based on the absolute value.In addition,under the premise of ensuring the quality of the fused image,the optimal NSCT structure is selected to minimize the performance requirements of the FPGA.Finally,the feasibility of the implementation of the fusion algorithm based on FPGA is analyzed,and a best implementation scheme of image fusion algorithm based on NSCT is formed.Based on the overall structural design of the NSCT-based image fusion algorithm,this paper uses Verilog HDL to describe the design with a fully synthesizable RTL level,and uses the synchronous design method to improve the system reliability.The main modules of this design include: NSCT decomposition module,image fusion module,image reconstruction module and so on.In order to verify the correctness of each module function,this paper combines the FPGA development software QuartusII 15.0 and Modelsim to simulate the various modules of the above design and the whole fusion system.In addition,subjective and objective evaluation methods are used to evaluate the quality of the fused image,and the system performance is analyzed from the aspects of FPGA resource occupancy and real-time performance.The experimental results show that the design can achieve the fusion of infrared image and visible light image with resolution of 640*480 on FPGA,and the delay of the system is small,which can realize real-time image fusion.