《基于红外图像的目标特征提取方法研究学士学位论文 理工大学》由会员分享,可在线阅读,更多相关《基于红外图像的目标特征提取方法研究学士学位论文 理工大学(48页珍藏版)》请在金锄头文库上搜索。
1、理工大学学士学位论文I摘 要红外图像目标特征提取是红外预警系统与成像制导的核心技术之一。红外系统的隐蔽性好,抗干扰能力强,作用距离远,搜索速率快,无镜面回波,可昼夜工作,这些优良特性使其越来越多地应用于军事和民用领域中。因此,如何提取鲜明的特征来提高红外图像目标的检测、识别性能以及整个红外系统的智能化水平一直是国内外学者的研究方向和焦点。本文旨在研究图像特征提取方法,使得能在分类识别中得到较高的识别率。首先,简要介绍了红外图像分析、传统提取方法分析,主要是 Hu 不变矩的理论,研究它在提取图像特征所表现出的特性。其次,针对传统提取方法的弊端,研究利用小波不变矩的特征提取方法。实验结果表明,采用
2、小波不变矩特征提取方法可获得较高的识别结果,从而也满足了实际系统的实时性要求。虽然在针对旋转图像时需要进行归一化,但这并不影响它在未来科技中的发展。关键词: 红外图像;特征提取;Hu 不变矩;小波不变矩理工大学学士学位论文IIAbstractFeature extraction of infrared images targets are one of the key technologies in infrared early warning and imaging guidance. Infrared systems have good concealment,strong anti-in
3、terference ability, far operating range and fast search speed. Besides, they have no mirror echo and can work day and night. All of these advantages make the mused in military or civilian fields more and more. Consequently, how to extract the distinct features, to improve the detection and recogniti
4、on of infrared images and thus to improve the intelligent level of the overall infrared systems, has been the research direction and focal point for the domestic and foreign scholars.This paper focuses on the methods of image feature extraction and getting higher recognition rate for the classificat
5、ion tasks.Firstly, briefly introduced the infrared image analysis, analysis of the traditional extraction methods, mainly Hu invariant moment theory, studies the properties exhibited by the image feature extraction. Secondly, aiming at the disadvantages of traditional extraction methods, feature ext
6、raction method based on wavelet invariant moments. The experimental results show that, using the wavelet moment invariant feature extraction method can get higher recognition results, which also meets the real-time requirements of actual system. Although the normalization in the rotated image need,
7、but this does not affect it in the future development of science and technology.Keywords: infrared picture; feature extraction; Hu invariant moments; wavelet invariant moments理工大学学士学位论文III目 录1 绪论.11.1 研究背景及意义.11.2 红外目标特征提取现状.11.2.1 特征提取的目的和原则.21.2.2 特征提取方法.21.3 论文的主要工作与章节安排.32 红外图像处理及特征提取理论研究.42.1 红
8、外图像的优缺点.42.2 二值化处理.42.2.1 红外图像直方图及其特点.52.2.2 阈值分割.72.3 常见的特征提取方法.82.4 多项式不变矩几何矩.93 基于 Hu 不变矩和小波矩的特征提取方法研究.133.1 Hu 不变矩特征提取方法.133.1.1 Hu 矩的不变性.133.1.2 Hu 不变矩的构造和流程图.153.1.3 Hu 不变矩实现程序.173.1.4 实验图片及分析结果.193.2 小波不变矩特征提取方法.203.2.1 小波矩.203.2.2 小波矩的构造和流程图.203.2.3 小波矩实现程序.253.2.4 实验图片及分析结果.274 仿真结果分析.294.1 红外图像预处理结果分析.294.2 异类目标的特征提取结果与分析.30理工大学学士学位论文IV4