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:
光电经纬仪虚拟测试仿真与评估技术研究
作者: 张勇
学位类别: 博士
答辩日期: 2009-06-05
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: 虚拟测试 ; 虚拟目标 ; 轨迹-姿态仿真 ; 图像合成 ; 目标识别 ; 效果评估 ; 模糊调节
其他题名: Simulation and Evaluation Technology of Virtual Tests for Optoelectric Theodolite
学位专业: 信号与信息处理
中文摘要: 针对光电跟踪测量技术快速发展过程中对系统及内部算法越来越迫切的测试与评估需求,搭建了虚拟测试评估平台体系结构,提出了一种基于虚拟环境的光电经纬仪虚拟仿真测试方法,建立了不同应用环境下系统及算法相关性能的有效测试评估方法,主要研究内容和结果如下。 针对虚拟测试场景仿真需求,主要研究测试场景目标仿真建模和场景图像仿真方法。通过对目标运动特征设计与仿真相对于具体应用的独立性分析,提出了基于运动学的虚拟目标运动轨迹和姿态的数据建模方法,摒弃了基于动力学的目标运动特征仿真方法。结合光电经纬仪的测量模型,得到了直接描述目标运动特征的可视化设计与仿真模型,实现了目标轨迹与姿态的独立建模与仿真,适应了OpenGL的成像特性。实验表明,三维场景中的目标仿真轨迹和姿态符合光电经纬仪的测试评估需求。 给出了虚拟测试场景中的经纬仪图像仿真方法,基于OpenGL的二维成像原理,导出了模拟光电经纬仪CCD相机的OpenGL虚拟相机模型,可以灵活的模拟出不同场景、不同目标和不同成像条件下光电经纬仪的理论成像结果。在实时仿真测试时,采用预先仿真出的目标图像经旋转和平移后叠加到预先仿真的背景图像上,并给出了目标图像叠加的平移量和旋转角的计算方法,从而逼真模拟经纬仪实际拍摄成像。实验表明,由以上仿真方法所生成的目标图像基本满足了当前测角误差、跟踪精度与姿态提取算法效果(或性能)的测试评估需求。 提出了一种面向弱小目标识别的非实时参数化图像仿真方法,弥补了实时成像的不足之处。探讨了目标图像特征及其影响因素与度量模型,采用虚拟目标模型,参数化设计其静态特性和运动特征,设计和配置背景特征及噪声模型,然后合成目标图像和背景图像,从而得到量化的理论仿真图像模型及数据。实验表明,该仿真方案所生成的合成图像有效拓展了非实时的自动目标识别(Automated Target Recognition,ATR)系统及其算法效果的检测与评估。 针对虚拟测试评估需求,主要研究基于上述仿真方案的系统及算法相关性能评估指标体系、不同应用条件下的评估方法。针对系统动态测角误差、跟踪精度和姿态提取算法检测,采用最大误差和均方根误差来衡量其效果。针对图像小目标的ATR系统及其算法效果评估,建立了适于不同应用环境的评估指标与评估方法。 将ATR算法评估分成图像预处理阶段(目标检测识别阶段)算法评估和目标跟踪识别阶段算法评估,对其识别效果的参照信息及其度量模型进行了设定。改进了Logistic回归模型,对识别算法正确识别率(the Probability of Detection-PD, )这一指标进行评估,以量化 随信噪比(Signal Noise Ratio,SNR)和目标大小的变化规律。采用了综合接受者工作特性(Summary Receiver Operation Characteristics,SROC)模型对 和虚警率(the False Alarm Rate-FAR, )两指标效果综合量化评估。修正了 指标的量纲,使其更适于计算和评估;改善了综合指标值,对算法的 , 进行综合效果评估,同时也获得算法的最佳工作点。针对目标跟踪识别阶段算法评估,另增加了跟踪稳定性(关联性能)和有效跟踪精度指标,采用关联矩阵法进行综合评估。采用了基于多目标的自由接受者工作特性(Free Receiver Operation Characteristics,FROC)模型,以单目标识别作为特例,对算法效果进行评估。实验表明,建立的评估方法能有效完成两阶段主要识别算法基于 和 的综合评估。 给出了基于图像的ATR系统(或主要算法组合)整体识别效果模糊综合评判方法,搭建了较完备的指标体系,确定了设置权重的直接权重法、W路径法和变权法;对样本获得条件、样本容量和假设检验模型进行了设定;界定了各指标值的评价等级统计区间;最后,探讨了模糊评判算法的一般模型及细化模型。实验表明,所建立的评估指标体系和评估方法能有效的对ATR系统整体识别效果进行评估,获得其在各外因水平下多种识别效果等级出现的概率和可能性。 针对虚拟测试平台下目标跟踪滤波算法的评估及应用,探讨了跟踪滤波算法的评估指标和评估方法,添加了噪声处理效果的辅助指标,引入协方差分析描述函数技术和谱分析方法,对滤波效果进行统计、分析;改进并给出了一种基于“当前”统计模型的分段模糊自适应跟踪方法。实验表明,采用该评估方法对改善前后的滤波算法效果进行分析比较,改善后的滤波算法取得了较好的跟踪效果。 利用上述可视化目标轨迹与姿态仿真技术、基于OpenGL的目标图像仿真技术、信息交互技术、面向弱小目标的参数化图像仿真技术和不同应用条件下的测试评估技术等,为光电经纬仪的研发拓展了多层次、多角度测试评估手段。通过实际测试评估应用,验证了虚拟测试仿真与评估系统及其关键技术的有效性。
英文摘要: With the rapid development of optoelectronic tracking and measurement technology, the test and evaluation for optoelectronic theodolite system and its inner algorithms are urgently demanded. An architecture of the virtual simulation testing platform is discussed. A virtual simulation testing method and a suit of evaluation technologies are proposed, based on the vary virtual reality and simulation environments. The main content and result are as follows. The simulation technology of the virtual testing scene is mainly involved with the target simulation modeling and the image simulation. The target motional character design is independent from real applications. A data modeling method for the target motional characters based on kinematics is proposed, which simplifies the method based on dynamics. Combined with the measurement model of optoelectronic theodolite, the simulation model directly describing the target motion is gotten, which makes the target track and pose independently modeled and simulated, and suitable for OpenGL imaging traits. Tests show that simulation tracks and pose in the testing scene are effective. OpenGL virtual camera is designed for theodolite imaging simulation, replacing the CCD camera. This virtual camera may simulate the ideal image sequences for different scenes, different targets and vary imaging conditions. Furthermore, the method synthesizing the beforehand simulated target image and the simulated background image in advance corresponding to the translating size and the rotating angle is proposed, which produces reliable theodolite images in real time. Tests show that simulation images produced by this method are effective for relative performance tests. A parameterized image simulation method for dim targets in not real time is proposed, which improves the real time simulation method. The characters of the target image,and the influence factors and the discribling model of them are discussed. The virtual target model is employed, which static characters and dynamic characters are quantitively designed and simulated. The background image mode is selected and its noise model is deployed, which synthesizes the former target’s image, and the testing image and it’s data are gotten and provided for the target recognition of algorithms. Tests show that simulation images produced by this method develops the tools for the evaluation of the automated target recognition (ATR)system and algorithms. The virtual testing evaluation technology is based on the above simulation methods and the virtual testing scene produced by them, and a suit of evaluation methods adaptable to vary application conditions are proposed, such as the evaluation of tracking measurement error, the track precision and the pose measurement. Also, the ATR evaluation items and methods are mainly established to adapt to vary environments. The performance evaluation of ATR algorithms is divided into the target detection recognition and the target tracking recognition evaluation. The key influence factors and the describing model of the recognition performance are discussed. The logistic model for describing the relationship between the Probability of the Detection (PD) and signal noise ratio(SNR) or target area(TA) is improved, which quantifies their relationship by the step or the back regression method. Summary receiver operation characteristics(SROC)model is employed to evaluate PD and false alarm ratio(FAR), which parameter’s dimension and evaluation items are improved. The summary performance is gotten by calculating the improved items. The relative tracking performance and the effective tracking precision are added to further evaluate the tracking recognition algorithms by the relative matrix method. Also, free receiver operation characteristics(FROC)model is employed for the single target recognition algorithms. Tests show that the above methods for PD and FAR evaluation are available. A fuzzy summary evaluation method for the ATR system or its algorithms grouping based on images is proposed. The evaluation items are selected and designed, which weights are calculated by the direct method, W-road method or the changing weight method. The testing conditions , the size and the hypothesis testing methods of the statistic swatch are discussed, which effectively tests the mean, the covariance dependency and the distribution of PD. The statistic ranges corresponding to evaluation ranks of these items are established. Finally, the simple model and the divisional model of the fuzzy evaluation algorithm are discussed. Tests show that the method is valuable for getting the probabilities of different performance rack of algorithms in vary environments. As for the evaluation and application of target tracking filter algorithms in virtual tests, the evaluation items and method of target tracking filter algorithms are discussed, and the noise process item is added. The covariance analysis describing function technique(CADET) and the frequency chart method are used to describe the process performance. Meanwhile, the adaptive maneuvering range of the “current” statistical model(CSM)is established and an improved fuzzy adaptive tracking algorithm is proposed. Tests show that the improved algorithm is more effective than the not one by this evaluation method. The visual simulation techniques, the OpenGL virtual simulation techniques, the information interactive techniques, the parameterized simulation method for dim target images and the evaluation techniques for different algorithms are proposed and employed, which can test and evaluate different levels and aspects of capabilities of optoelectronic theodolite Tests show that the virtual testing and evaluation system and the key technologies are effective and available.
语种: 中文
内容类型: 学位论文
URI标识: http://ir.ioe.ac.cn/handle/181551/350
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_200618015127026张勇_paper.pdf(4134KB)----限制开放View 联系获取全文

Recommended Citation:
张勇. 光电经纬仪虚拟测试仿真与评估技术研究[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
[张勇]'s Articles
CSDL cross search
Similar articles in CSDL Cross Search
[张勇]‘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_200618015127026张勇_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