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:
移动背景下目标检测与跟踪技术研究
作者: 卢宇
学位类别: 博士
答辩日期: 2009-06-04
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: 吴钦章
关键词: 目标检测 ; 目标跟踪 ; 电子稳像 ; 移动背景 ; 空地成像
其他题名: Target Detecting and Tracking in the Moving Background
学位专业: 信号与信息处理
中文摘要: 移动背景下的目标检测与跟踪技术研究探讨了机载应用场合中对地物背景下可视目标的检测和跟踪技术。成像设备安装在飞行器上,导致成像设备高速运动,反映在图像上表现为背景移动;成像设备与地面或低空目标之间距离较远,成像过程易受气象条件影响,导致视频图像序列成像质量有限。在对目标检测和跟踪的基本理论、算法和技术进行了深入分析的基础上,针对移动背景和空地成像的约束条件提出了相应的解决办法。 首先,针对背景高速移动的特性,对电子稳像技术进行了研究,提出了基于纵横序列的全局运动参数快速提取算法。所谓纵横序列,是指基准帧中的特征区域呈同一水平位置,或同一垂直位置排列。利用这种先验信息可以简化和重构全局运动参数的提取流程。这种改变不仅涉及到全局运动参数提取算法自身,也涉及特征区域的选取算法和匹配算法。实验表明在不影响整体精度的前提下,本文提出的快速算法具有突出的高效性,为工程应用打下了基础。 其次,针对可能出现的电子稳像和背景配准不够充分的情况,提出了基于迭代的Lucas-Kanade算法。通过层内迭代和层间迭代,增强了光流场的计算精度,提高了目标检测对移动背景的适应性;针对实际工程应用场景中通常不能事先确定视场中有无目标及其目标数量的问题,提出了基于弹性分组逻辑的模糊聚类法分析光流场。该方法通过建立目标稳定性模型和目标分裂模型,自适应地分割和定位目标。实验表明这套基于光流的检测方法无论是运动场的计算精度,还是在运动场目标提取方面都明显地提高了对背景移动和目标定位的鲁棒性。 第三,将人眼视觉系统(Human Visual System, HVS)的某些特性与粒子滤波技术框架相结合,提出了一种基于HVS特性的多特征跟踪技术。在直觉层次上,建立运动特征距离模型;在专注层次上,建立SVD不变性距离模型,同时采用灰度直方图距离模型;多个特征的组织统一于粒子滤波框架下,共同决策跟踪判决结果;按照HVS思想建立反馈机制调节各层次信息对跟踪融合的贡献度。实验表明这种方法在跟踪过程中具有智能特性,能够在复杂背景场合和目标发生形变时稳定地跟踪目标。 最后,研究了移动背景下目标检测和跟踪系统实现的软硬件技术。采用了接口技术的思想,设计了系统的整体框架,屏蔽了底层硬件特性和算法执行细节,保障了上层控制逻辑的稳定性,增强了软件系统的可移植性、可重用性和可扩展性。 总之,论文针对移动背景和空地成像的约束条件,结合工程实际,对机载目标检测和跟踪系统进行了广泛而深入的研究,提出了一套针对性很强的思路和算法。
英文摘要: The visible target detecting and tracking techniques in the moving background for air-to-ground applications are discussed. The imaging equipment is fixed in the aircraft, which leads to its high moving speed and further is reflected as the background movement of images, the larger distance between the imaging equipment and the ground or low-altitude target. In the imaging process, the influence of the weather condition results in the precision limits of the image sequences. On the basis of analyzing the foundational theories, algorithms, techniques and solutions aiming at both the moving background and the air-to-ground imaging are proposed. First of all, in view of the high moving speed of the background, the electronic image stability (EIS) is analyzed. A fast global movement parameters pick-up algorithm based on cross features regions is proposed. The cross refers to the features in the reference frame distribution in the same row or same column. Making use of such prior information, computation of the global movement parameters’ can be reduced and reconstructed. This modification involves not only the parameters pick-up algorithm itself but also the features’ selection and matching. Experiments illustrate that the proposed algorithm provides the prominent efficiency while guaranteeing the precision, which contributes to engineering applications. Secondly, in view of the possible fact that the EIS and background registration is not exact, an iterative Lucas-Kanade algorithm is proposed. In virtue of the pyramid structure, the iteration process includes in-layer iteration and out-layer iteration. The precision of the optical flow field is improved. The algorithm can be suitable for the condition that the background is moving. In view of the fact that the target number is not easily determined in the practical detection environment, an improved fuzzy clustering method based on elastic grouping logic is proposed. This method features in establishing the target stability model and its division model to partition and locate the targets adaptively. Experiments show this suit of detection solution based on optical flow field improves the robustness of the moving background and targets locating in the processes of both the optical field computing and abstracting. Thirdly, certain characters of human visual system (HVS) are combined with the framework of particle filter technique, and a multi-cues tracking method based on HVS characters is proposed. For the instinctive level, the motion cue model is established; for the dedicated level, the singular value decomposition (SVD) invariance model is established; and the grey histogram is also utilized; all cues are organized in the freamework of the particle filter to determine the tracking result; a feedback mechanism is established to adjust the contribution each HVS cue. Experiments show this method is smart and able to stably track the target under the complex background even thought the target is deformed. Finally, the implement techniques about both the hardware and the software are analyzed. Employing the interface technique, the whole software framework is designed, which shields the hardware differences and the algorithm detail. So it guarantees the stability of upper control logics and enhances the transplantation, reuse, and expansibility of the software system. In conclusion, in view of the two constrains as well as moving background and air-to-ground sence, combined with the project application, this dissertation researches into the target detecting and tracking system mounted in aircrafts and proposes a suit of ideas and algorithms with strong pertinences
语种: 中文
内容类型: 学位论文
URI标识: http://ir.ioe.ac.cn/handle/181551/349
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_200618015127024卢宇_paper.pdf(3931KB)----限制开放View 联系获取全文

Recommended Citation:
卢宇. 移动背景下目标检测与跟踪技术研究[D]. 光电技术研究所. 中国科学院光电技术研究所. 2009.
Service
Recommend this item
Sava as my favorate item
Show this item's statistics
Export Endnote File
Google Scholar
Similar articles in Google Scholar
[卢宇]'s Articles
CSDL cross search
Similar articles in CSDL Cross Search
[卢宇]‘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_200618015127024卢宇_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