Research On Enhancement And Zooming Of Infrared Video Signals

With the development of modern infrared technology, Infrared imaging systems are developed rapidly in commercial fields (such as infrared monitoring) and military fields (such as infrared guidance). Because the infrared images that are low contrast don't have as much scales of gray as visible images, it is difficult to identify target. Base on the research subject of certain type of missile IRIGS, this dissertation use the algorithms of infrared image enhancement and zooming to realize the infrared video signal acquisition and control and real-time processing in FPGA.Firstly, this dissertation analyzes the characteristics of the infrared image. On the basis of real-time realization, this dissertation researches on the infrared image enhancement methods of gray transformation and spatial filtering and proposes a improved algorithm for adaptive piecewise-linear transformation, then combining it and spatial enhancement methods for the ultimately algorithm. In order to making this infrared image enhancement system adaptive, a new method that based on the image histogram analysis and select automatically the most appropriate algorithm for infrared image is proposed.Then this dissertation analyzes the nearest neighbor interpolation, bilinear interpolation, bicubic interpolation and pixels exchange interpolation. At the real-time requirements, an improved algorithm of pixels exchange interpolation is selected. After simulation, this method has a balance of the complexity and the effect.Finally, this dissertation describes a video signal-processing system, which based on FPGA+SRAM architecture and including: video acquisition module, I/O frame storage module, video processing module and video output module. Then studying on the functions and applications of video decoder SAA7115 in the acquisition module, for example, the PAL video signal decoding, window acquisition, video scaling. In the frame memory module, the functions of Video Capture control, image enhancement algorithms and video output control is wrote by Verilog HDL. In image processing module, HDL does the algorithms of infrared image enhancement and zooming. The results of QuartusII simulation show that the algorithms of image enhancement and zooming are good and can get real-time request.