Research And Implementation Of Two-color Infrared Image Fusion Processing Technology

At present,single-band imaging is mainly used for infrared imaging.But the scene information obtained by single-band imaging is limited,which cannot meet the needs of multiple scenarios.In order to obtain more comprehensive scene information,multi-band infrared detection and image fusion technologies become the focus of current research in the infrared field.The main contents of this article include:1)Aiming at the shortcoming of the infrared image with low contrast,a block-based histogram equalization algorithm is proposed.This algorithm performs histogram statistics and accumulation on the divided images,and proposed a new block mapping method to avoid blocking effects.Compared with the commonly used global platform histogram,this algorithm can improve the contrast of infrared images effectively.2)Fusion of dual-band images.The fusion algorithms selected in this paper are Laplace pyramid fusion and TNO fusion.Aiming at the shortcomings of Laplacian pyramid fusion image with insufficient details and poor overall contrast,a pyramid fusion algorithm based on enhancement is proposed,which improves the amount of detail information and the contrast of the fused image effectively.In the TNO algorithm,large area color makes the "bottom of the pot" effect of the fused image prominent.Aiming at this phenomenon,a correction-based TNO fusion algorithm is proposed.This algorithm displays color only at locations with large differences in dual-band gray levels,and reduces the "bottom of the pot" effect of the fused image.3)Completed system design,debugging,software architecture.Realized the algorithm.In terms of hardware,FPGA is used as the core of fusion circuit design.In terms of software,the infrared image processing architecture based on Qsys platform is built.This architecture realized the hardwareization,modularization(IP core)and module detachability of the image processing algorithms.The Nios II processor was used to uniformly schedule and manage all IPs to achieve efficient and real-time operation of the image processing architecture.From the final output fusion image,it can be concluded that the contrast of the pyramid fusion image is improved significantly,the bottom effect of TNO fusion image is weakened,and the overall visual effect of the image is improved.