中国科学院光电技术研究所机构知识库
Advanced  
IOE OpenIR  > 光电技术研究所博硕士论文  > 学位论文
题名:
目标红外辐射特性测量关键技术研究
作者: 苏赋
学位类别: 博士
答辩日期: 2007-06-11
授予单位: 中国科学院光电技术研究所
授予地点: 光电技术研究所
导师: 杨文淑
关键词: 辐射测量 ; 作用距离 ; 非均匀性校正 ; 图像分割
其他题名: Research on the key Technique of Calculating Target’s Radiance in Infrared Image
学位专业: 信号与信息处理
中文摘要: 根据红外热成像系统获得的目标红外图像进行目标的红外辐射特性测量已成为国内外的研究热点之一。本文围绕红外热成像系统和目标红外辐射特性测量中的关键技术,展开了深入的研究,研究内容与结果包括以下几个方面: 对目标与背景的红外辐射和辐射大气传输和凝视型红外热成像系统作用距离模型进行分析,提出了基于噪声等效温差的作用距离模型。该模型从红外热成像系统的信噪比和噪声等效温差(Noise Equivalent Temperature Difference, NETD)定义出发,综合考虑背景辐射和目标成像弥散的影响。 针对根据红外图像进行目标红外辐射测量,总结了实际测量中的关键技术。主要包括红外焦平面阵列的非均匀性校正、定标、大气传输计算、红外图像处理和目标红外辐射计算。为了将目标从背景中提取出来,只有经校正后,均匀性好的图像中才能准确区分出目标。因此,均匀性效果直接影响到提取目标的精度,故本文主要进行了红外焦平面阵列的非均匀性校正和红外图像处理方面研究。 基于红外焦平面阵列的非均匀性特性,研究了非均匀性产生的原因和近年来讨论较多、相对较为成熟的红外焦平面非均匀性校正算法,并进行了分析和比较。在此基础上提出了基于图像梯度的神经网络红外焦平面非均匀校正算法,从理论上研究了神经网络算法校正系数自适应的原理,从而提出了在传统人工神经网络校正算法基础上通过修改像素的期望输出和学习速度,具有基于图像梯度和自适应学习速度的新的红外焦平面非均匀校正算法。仿真和试验结果表明算法收敛速度和校正效果都优于传统的神经网络校正算法。 在红外图像处理方面主要着重于图像分割的研究,图像分割的目的是区分出背景和目标,为后续目标能量的统计打下基础。本文提出了几种新的图像分割技术。基于标准差的红外小目标分割算法对图像进行基于标准差区域划分,将图像划分为边缘区和平滑区,然后再利用Otsu方法对边缘区内像素点进行分类,最后对分类后的二值图像进行孔洞填充完成对红外图像的分割。红外小目标小波多尺度相关检测方法,考虑到边缘和噪声的Lipschitz指数不同,在不同尺度上小波变换模有不同传播特性,将传统的基于小波变换模极大值的小波多尺度边缘检测和小波变换尺度间相关性的特点相结合。基于MAS小波红外小目标检测方法,通过二进MAS小波对图像进行多尺度分析,根据边缘和噪声的小波变换模在各尺度下的不同传播特性,计算相邻尺度的小波变换模相关量用来增强信号,抑制噪声,比较归一化小波变化模相关系数和小波变换模区分边缘和噪声,进行红外小目标的检测。实验结果证明,本文提出的几种方法相对传统的算法更加能够有效的和准确的分割目标。 本论文针对根据红外热成像系统获得的目标红外图像进行目标的红外辐射特性测量中的难点和关键技术,进行了一系列理论、方法和工程应用的探索研究,提出了一些针对性很强的思路和方案,为以后开展相关工作奠定了良好的基础。
英文摘要: The technique of calculating target’s radiance in infrared image gained by infrared imaging system has given more and more interests for researchers. Infrared imaging system and the key techniques of calculating target’s radiance are researched in this PhD dissertation, and the main contents and results are as follows: Studying on the radiance of target and background and the model of operating range of a staring IR imaging system, a new equation of operating range expressed by NETD is deduced, considering the background radiation and image dispersion. With respect to calculating target’s radiance in infrared image, the main technologies of calculator are summarized, which included the non-uniformity correction for infrared focal plane array, calibration, atmosphere modification calculation, image processing and target and background selection and radiance calculation. This dissertation mainly studies the non-uniformity correction for infrared focal plane array and image processing. In the field of the non-uniformity correction for infrared focal plane array, the attribution of non-uniformity and some mature arithmetic of non-uniformity correction are discussed. An image gradient-based neural network algorithm is present which uses the optimization techniques like improvement in desired value of pending pixel and adaptive learning rate. The new algorithm has been tested with real infrared data sequence and has the advantage of high convergence speed and improvement of correction result than the original neural network one. The PhD dissertation discusses some new algorithms of image segmentation especially for small target. A image segmentation algorithm based on local standard deviation uses the maximum between-cluster variance as a rule to segment infrared image into edge and inner regions to implement edge detection. Since the edge region consists of the edge of target and that of background, the contour line is wide. The inner region obtained in edge detection is rejected and the edge region is retained. Once more segmentation using traditional Otsu method for the edge region bring forth the edge of target and that of background. Finally, holes are filled in the image marked the target edge. Based on wavelet scale correlations a new algorithm of infrared small target detection is presented. The discrete dyadic wavelet transform is employed to produce the multi-scale representation of image, correlations of big scales' modulus are calculated based on the difference characterization of edges and noise by Lipschitz exponents and a coarse edge map is then scavenged, logical-AND operation is performed between the coarse edge map and the modulus map of small scale and the maxima of the wavelet transform modulus are checked out, at last, infrared small target is detected. The algorithm based on MAS wavelet transform is also discussed. The discrete dyadic MAS wavelet transform is employed to produce the multi-scale representation of image, correlations modulus are calculated based on the difference characterization of edges and noise by Lipschitz exponents and then the edge and the noise are separated, small target and the edges of background and clouds are separated based on the different singularity of them of the same scale. Aiming to the key and difficult technique of calculating target’s radiance in infrared image, this Ph D dissertation includes research of theories, techniques in the area. The work of this dissertation has great contribution and important reference to the future development of this area.
语种: 中文
内容类型: 学位论文
URI标识: http://ir.ioe.ac.cn/handle/181551/226
Appears in Collections:光电技术研究所博硕士论文_学位论文

Files in This Item:
File Name/ File Size Content Type Version Access License
10001_200418015127020苏赋_paper.pdf(1260KB)----限制开放View 联系获取全文

Recommended Citation:
苏赋. 目标红外辐射特性测量关键技术研究[D]. 光电技术研究所. 中国科学院光电技术研究所. 2007.
Service
Recommend this item
Sava as my favorate item
Show this item's statistics
Export Endnote File
Google Scholar
Similar articles in Google Scholar
[苏赋]'s Articles
CSDL cross search
Similar articles in CSDL Cross Search
[苏赋]‘s Articles
Related Copyright Policies
Null
Social Bookmarking
Add to CiteULike Add to Connotea Add to Del.icio.us Add to Digg Add to Reddit
文件名: 10001_200418015127020苏赋_paper.pdf
格式: Adobe PDF
此文件暂不支持浏览
所有评论 (0)
暂无评论
 
评注功能仅针对注册用户开放,请您登录
您对该条目有什么异议,请填写以下表单,管理员会尽快联系您。
内 容:
Email:  *
单位:
验证码:   刷新
您在IR的使用过程中有什么好的想法或者建议可以反馈给我们。
标 题:
 *
内 容:
Email:  *
验证码:   刷新

Items in IR are protected by copyright, with all rights reserved, unless otherwise indicated.

 

 

Valid XHTML 1.0!
Copyright © 2007-2016  中国科学院光电技术研究所 - Feedback
Powered by CSpace