多光谱融合前端图像实时降噪处理系统的研究

【作者】 杨晋伟；

【导师】 常本康；

【作者基本信息】 南京理工大学 ， 光学工程， 2009， 硕士

【摘要】 多光谱图像融合技术无论是在军用还是民用领域都有着强劲的发展势头和广泛的市场应用前景。微光与红外图像的融合是夜视技术领域最为朝阳的一项新技术。然而,作为图像融合技术中的关键技术问题之一的图像融合前端图像预处理技术一直未能得到比较理想的解决。到目前为止,市场上还没有出现针对图像融合前端降噪优化预处理的处理系统。文章就“多光谱融合前端图像实时降噪处理系统的研究”这一课题,研制了一款小型化的、性能稳定的、针对微光与红外图像融合前端降噪预处理的实时图像处理器。文章介绍了图像融合技术和图像融合预处理技术的基本发展;对图像的主客观评价方法进行了研究分析;针对微光和红外图像的各自特点提出了优化预处理方案;对各种图像优化处理的算法进行了分析和研究,特别针对图像的实时降噪进行了多模自适应递归降噪算法的研究;探索了适合于实时图像优化预处理的算法方案;完成了以FPGA为处理核心的小型化实时图像处理系统的整体解决方案;按照系统方案设计了硬件系统的各功能模块电路,并完成了小型化硬件系统的PCB设计;完成了基于FPGA的软件系统方案设计;以逻辑系统设计的方法为主并结合SOPC技术,完成了各软件功能模块的设计,成功搭建了图像预处理的算法实现平台;针对微光图像的预处理完成了多模自适应递归算法和灰度变换算法的软件设计;完成了整个软硬件系统的功能调试,并进行了实际环境中的图像处理实验和实验结果分析。文章最终研制出了能够基本满足图像融合系统中前端图像实时降噪预处理的小型化处理器,并为更多功能和更有效的算法实现提供了成熟的平台。更多还原

【Abstract】 The technology of multi-spectrum image fusion has a very good development tendency and a widespread market application prospect in both military and civilian domain. The fusion of the low level light and the infrared image is a very popular technology in the night vision domain. However, as one of the key technologies in the image fusion domain, the image pretreatment before the fusion still has no ideal solution. So far, there has not been any processing system to the optimal pretreatment of front-end image noise reduction in the market. The paper focus on the topic of "The research of the real-time image noise reduction system in front of the multi-spectrum fusion", and a small stable real-time image processor which is used to do noise reducing pretreatment in front of the fusion system of the low level light and the infrared image is developed.In the paper, the development of the image fusion technology and pretreatment of the image fusion are introduced; Some image subjective and objective assessment methods are analyzed; The project of the pretreatment on the low level light and the infrared image are proposed respectively; Several algorithms to the optimal pretreatment of various kinds of images are analyzed, and the multimode auto-adapted recursion algorithm which is aimed at the image real-time noise reduction is studied specially; The project of the choice of algorithms for the real-time image optimal pretreatment is explored; The overall solution of the real-time image processing system which is based on FPGA is completed; According to the system project, various functional modules of the hardware system are designed, and the PCB design is completed; The project of the software system design based on the FPGA is completed; Various software functional modules are completed by the logical system design method Primarily and the SOPC technology, and the platform for the realizing of image pretreatment algorithms is built; The software designs of the multimode auto-adapted recursion algorithm and the gradation transformation algorithm which are chosen for the low level light image pretreatment are completed; The hardware and software system debugging, the test of image processing in the actual environments and the result analysis are completed.Finally, a real-time noise reduction pretreatment processor which is able to satisfy the image fusion system basically and provide a mature platform for multi-purpose and the more effective algorithm realization is developed.更多还原