Infrared And Visible Light Image Fusion System Research Based On DSP

In modern times, image fusion as a new cross subject in the new and comprehensive fields, is more and more applied in the network video monitoring, target detection, military operations, and medical equipment. Image fusion covers computer artificial intelligence, A/D conversion, digital image processing, computer graphics and image processing, and machine vision, etc. With more and more advanced hardware equipment, the demand of record information on the spot becomes higher and higher. Infrared images are from the infrared sensors can well identify target information and have good penetration, hidden target information can be detected in the military operations, but background information is a little; Visual image is shot by the visual sensor and has abundant background information, but hidden target information is not detected, it is influenced by environment. Infrared image and visual image are fused to achieve the better visual effect, the basis for subsequent image processing is provided.The basic principle of multi-scale and multi-direction image fusion algorithm, which is hot research in recent years, this image fusion algorithm is studied and understood. The imaging mechanism of different sensors and acquired images are fused. The fused algorithm is tested by DM642 which is embedded development platform, and used for the engineering practice. The following aspects are researched in this paper:(1) The background, research status and application direction of image fusion are deeply and comprehensively discussed.(2) The theory of image fusion is explained. The process, level, theory framework, subjective and objective evaluation of image fusion are introduced in detail. Mainly theory methods of image fusion are listed.(3) Image fusion algorithm based on NSCT is proposed and deeply studied. Infrared and visible light images by strict registered are correspondingly fused under its theory framework. Then the optimal fusion strategy is obtained.(4) Programs are implemented on the hardware development system TMS320DM642. The image pixel level fusion simulation by using high-speed computing power of TMS320DM642.