基于红外偏振成像的舰船目标识别及其系统实现

发布时间：2018-03-13 06:08

本文选题：红外偏振　切入点：图像处理　出处：《烟台大学》2017年硕士论文　论文类型：学位论文

【摘要】：舰船红外偏振成像目标识别技术是红外制导技术中十分重要的一个方面。现如今,人们越来越重视海面舰船目标识别技术的应用。到目前,针对普通红外图像的处理,已经提出了很多目标识别算法,处理效果也相对不错,但是针对能够有效抑制干扰的红外偏振成像制导的目标识别和系统实现还是有些许不足,所以更优化的红外偏振图像目标识别处理算法,仍是现在研究的关键所在。为此,本文对近岸舰船红外偏振图像目标识别的处理算法进行了探究,针对偏振图像的特点提出了一套完整的识别算法并通过硬件进行了系统实现。论文首先从红外偏振成像技术的角度出发,对该技术国内外发展现状进行了总结,分析了近岸舰船目标红外偏振成像的特点,根据这些特点探究了红外偏振图像增强技术,并对红外偏振舰船图像的图像预处理算法讨论;接着对红外偏振舰船图像的目标识别与目标跟踪算法进行了研究,最后进行了系统实现和验证。本文设计的红外舰船目标识别相关算法,包括基于支持向量机的识别技术和基于模板匹配的识别算法,首先通过MATLAB软件进行仿真验证,实验结果表明该方法正确识别到了舰船目标。随后在项目给定的应达到的技术指标和参数的基础上研究了系统的软件设计和硬件架构。最后利用TMS320DM642芯片进行了算法实现的系统验证。实验结果表明该方法正确识别到了舰船目标,大大减少了计算量,提高了识别精度,缩减了识别时间,实现了项目要求的参数要求,达到了项目的技术指标。
[Abstract]:Ship infrared polarization imaging target recognition technology is a very important aspect of infrared guidance technology. Nowadays, people pay more and more attention to the application of ship target recognition technology. At present, the common processing of the infrared image, a target recognition algorithm has been proposed, the treatment effect is relatively good, but for the target recognition and system infrared polarization imaging guidance can effectively suppress interference or achieve some shortcomings, so more optimization of the infrared polarization image target recognition algorithm, the key is still the research now. Therefore, the processing algorithm of target recognition inshore ship infrared polarization image were studied, according to the characteristics of polarization image and put forward a complete set of the recognition algorithm and through hardware implemented. Firstly, from the perspective of infrared polarization imaging technology, the technology of domestic Development status quo are summarized, analyzed the characteristics of the infrared polarization imaging of ship targets near shore, these features on infrared polarization image enhancement technology based on image preprocessing algorithm, and discuss the infrared polarization image of a ship; then the infrared polarization image of ship target recognition and target tracking algorithms are studied, finally realized and verify the algorithm for infrared ship target recognition. In this paper, which is based on support vector machine recognition and recognition algorithm based on template matching, simulation by MATLAB software, the experimental results show that the method is correct and recognition to the ship target. Then based on technical indicators and parameters should be given on the project research on software design and hardware architecture. Finally, using TMS320DM642 chip for system verification. Experimental results of algorithm The results show that the method correctly identifies the ship target, greatly reduces the amount of computation, improves the recognition accuracy, reduces the recognition time, achieves the requirements of the project parameters, and achieves the technical specifications of the project.

【学位授予单位】：烟台大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：E925;TP391.41

【参考文献】