FPGA-based Dual-band Infrared Image Fusion Real-time Processing System

Infrared imaging system through the scene thermal radiation imaging,with high contrast,long detection range,environmental adaptability and other characteristics of the commonly used infrared imaging imaging wavelength band for the long wave infrared and medium wave infrared.For the same scene,long-wave imaging has the characteristics of high gray level and clear detail,while the middle wave imaging has the characteristics of high contrast and clear target,and combines the advantages of long-wave and mid-wave images complementarily into one image to improve the system's observational ability.The image fusion real-time processing system studied in this paper is based on FPGA,using Qsys to build the system architecture,adding NIOS processor to control the system and improving the flexibility of the system;and designing Avalon bus-compliant data stream for different image data streams interface.Based on the traditional image fusion algorithms,the Laplacian pyramid fusion algorithm is improved according to the characteristics of infrared long-wave and infrared medium-wave images by adding the gradient-based adaptive weights,So that the fused image can better preserve the edge information and contrast of the image.In the aspect of color fusion algorithm,TNO fusion is carried out after the brightness and contrast are transferred,which improves the shortcoming of traditional TNO algorithm because of the monotonous and low contrast background of the fused image due to the overall brightness of the long wave image being higher than that of the middle wave image.This kind of algorithm has completed the FPGA implementation.A complete fusion system includes image input and output,image preprocessing,image fusion and image postprocessing.In the first chapter,the characteristics of long-wave imaging and mid-wave imaging,the detector of imaging system and the hardware system are briefly introduced.The second chapter introduces the common image fusion algorithm and the improved image fusion algorithm based on the characteristics of dual-band infrared images.The third Chapter four introduces the software architecture of the system and the realization of the fusion algorithm in detail.The fourth chapter introduces the image preprocessing and post-processing algorithms and implementation of the system.The fifth chapter introduces the hardware design of the system and FPGA driver of video input and output.