云层背景下目标多特征信息融合及跟踪策略研究
|
摘要
靶场光电跟踪测量系统利用光学成像技术跟踪测量空间的飞行目标，以获取目标的高精度测量数据，为武器、航天试验鉴定及故障分析提供重要依据的光电仪器。近年来，新的试验任务需求和靶场建设发展需要对靶场光电跟踪测量系统连续稳定跟踪目标的能力提出更高要求。然而，大量外场试验结果表明当前靶场光电跟踪测量系统的目标跟踪技术存在着对天气条件过于依赖的缺陷。尤其是在晴空中夹杂许多云层的情况下，一方面极易将云层错误识别成目标导致跟踪中断；另一方面在目标被遮挡后缺乏高效的搜索再跟踪能力，大多数需要花费较长时间并且依赖外界引导数据及人员的判决操作进行目标搜索再跟踪。由于晴空夹杂云层是靶场试验任务中常见的气象条件，为了避免局部云层的影响而错失后续的测量数据甚至是关键段的珍贵数据，极有必要研究并解决云层背景下目标跟踪易中断的问题。
本文以靶场光电跟踪测量系统为研究平台，以提高云层背景下目标跟踪准确性和平稳性为目的，围绕当前云层背景下目标跟踪方法存在的不足之处，开展目标识别、目标跟踪及目标搜索相关关键技术的研究。在方法具有实用价值的指导思想下，提出靶场光电跟踪测量系统在云层背景下稳定跟踪的实现方案。全文完成的主要工作归纳如下：
研究了云层背景下疑似目标区域分割技术。针对现有的单一图像分割技术难以实现云层背景下疑似目标区域的有效分割，确立了依据跟踪模式特点采用相适应的图像分割技术思路。将目标跟踪过程划分成稳定跟踪模式和主动搜索模式。在稳定跟踪模式下，依据目标大致位置区域已知的特点，采用改进后的最大类间方差阈值分割方法进行快速图像分割；在主动搜索模式下，依据全视场分割的要求，提出了基于视觉注意机制的图像分割方法实现精准的疑似目标区域分割，通过算例验证了方法的可行性。
研究了目标多特征信息融合识别技术。针对目标周边存在云层或被云层遮挡时，采用单一的灰度特征易将云层错误识别成目标的问题，提出了基于神经网络和证据理论两级联合的多特征信息融合目标识别方法。首先提取分割后疑似目标区域的多种特征信息，对每种特征信息通过相应神经网络进行特征识别，再利用证据理论对多个神经网络的识别结果融合判决。该方法一方面利用了目标多特征信息降低目标误识别的概率，另一方面两级联合的信息融合模型体现出良好的可扩展性，能方便地根据不同情形、不同对象增减改变特征融合信息源，具有很强实用价值。实验结果证明该方法能准确辨别出云层和目标。
研究了目标运动状态估计反馈跟踪技术。针对传统的目标脱靶量直接反馈跟踪方式难以适应云层背景下目标跟踪需求的问题，深入研究了采用目标运动状态估计反馈的跟踪方法及实现细节。分析了不同应用情形下目标运动状态估计的可观测性问题，探讨了卡尔曼滤波、不敏卡尔曼滤波及交互多模型滤波三种滤波估计方法的应用改进，给出了实际应用时目标运动模型建立、数据时空对齐、数据关联及状态估计等实现过程。实验结果表明该方法弥补了直接反馈跟踪方式的跟踪性能严重依赖于测量数据质量的缺陷，提高了云层背景下目标跟踪稳定性。
研究了运动目标主动搜索策略。针对目标跟踪过程中受到云层遮挡而短暂丢失目标的情形，提出了基于最大概率区间的目标主动搜索方法；针对长时间丢失目标且缺乏外界引导数据的情形，提出了在拦截点处采用警戒线目标搜索方法来提高搜索效果，同时针对警戒线搜索过程中目标成像会出现运动模糊的情况，给出了运动模糊恢复方法。实验结果表明主动搜索增大了目标有效搜索区间，可以大幅提高目标搜索成功概率。
综上，本课题通过对靶场光电跟踪测量系统现有的目标跟踪技术研究和改进，提出并验证了在云层背景下工作时目标多特征信息融合及稳定跟踪的相关关键技术，该技术的意义在于解决了云层背景下目标跟踪易中断的问题，弥补了当前靶场光电跟踪测量系统的目标跟踪技术对天气条件过于依赖的不足，在目标跟踪测量领域具有一定的实用价值。
A range photoelectronic tracking measurement system is the use of opticalimaging techniques tracking measurement space flight target, in order to get highprecision measurement data of the target, and provide important basis for testevaluation and fault analysis of weapons and spaceflight. In recent years, the needsof new test mission and range construction development put forward higherrequirements for range photoelectric tracking measurement system to continuouslyand stably track target. However, large numbers of field test results show that thecurrent target tracking technology of electro-optical tracking measurement systemexists defects that it relied too much on the weather conditions, especially theinclusion of many clouds in the clear sky. On the one hand, it is easy to misidentifyclouds into the target which will leads to tracking interruption; on the other hand, itlacks the ability of effective searching and tracking again when the target getcovered, most of the time it takes a long time and rely on external guide data andpersonal judgment operation to search and track target again. The clear sky withclouds is a common meteorological condition in shooting range test task. To avoidmissing out on the follow-up measurement data and even the key period of valuabledata for the influence of partial clouds, it is necessary to study and solve the problemof the easily interruption with tracking under the background of clouds.
The paper is based on the range photoelectric tracking measurement system asthe research platform, on the purpose of improving the target tracking accuracy andstability as well as around the deficiency of current target tracking method under thebackground of the cloud, to carry out the relevant key technologies research of targetidentification, target tracking and target search. With the ideology of practicality, thestable tracking scheme is proposed under the background of clouds.
The main works accomplished in the paper summarized as follows:
The suspected target region segmentation technology is researched under thebackground of the clouds. Image segmentation can effectively reduce the operationcosts of subsequent target identification. In view of the existing single imagesegmentation technique is difficult to realize effective segmentation suspected targetarea under the background of clouds, an idea is established that image segmentationtechnology is adopted flexibly according to the characteristics of the tracking mode.The target tracking of range photoelectric tracking measurement system is composedwith stable tracking mode and active searching mode. In stable tracking mode,beacause the rough target location is known, the improved maxinum between-clustervariance threshold segmentation is adopted for fast image segmentation. In activesearching mode, due to the requirements of the full view-field segmentation, theimage segmentation method based on visual attention mechanism is proposed torealize precise suspected target area segmentation. The experimental results showthat the method is feasible.
The target multi-feature information fusion recognition technology is researched.If the target is surrounded or covered by clouds, the cloud is easily identified intotarget through the single gray characteristics. Thus, the multi-feature informationfusion recognition method is proposed basing on two-level jointing neutral networksand evidence theory. Various feature information of the segmented suspecting targetarea is extracted firstly, and then the corresponding neutral network is used to featurerecognize. Afterwards, evidence theory is utilized to incorporate multiple neuralnetwork outputs and make the synthetical decisions. On the one hand, the target multi-feature information is used to improve recognition accuracy. On the other hand,the two-level joint information fusion model shows well extensible. It has stronglypractical value that it can easily increase or decrease feature fusion informationsource according to different situations or different object. The experimental resultsshow the clouds and the target can be accurately identified through the method.
The target motion state estimation feedback tracking technology is researched.Because the traditional way of direct feedback tracking target is difficult to meet therequirement of target tracking under clouds background, the target motion stateestimation feedback tracking technology is studyed deeply in the range photoelectrictracking measurement system. The observability problem is analyzed in differentapplication situations. Kalman filter, unscented kalman filter and interactive multiplemodel filter are discussed and improved for the application. The target motion model,data space and time alignment, data association and state estimation are achieved.The experimental results show that this method makes up for defects that the targettracking performance of direct feedback seriously dependent on the measuring dataquality, and improves the stability of target tracking in the cloud backgrounds.
The active searching strategy of moving target is studied. It puts forward the targetactive searching method based on the maximum probability areas in the situation ofcloud-covering and temporary loss of target. In order to speed up the searchingresults for long losting goals and lack of external guide data, it puts forward thewarning line target searching method at the intercepting point. Because the imagemotion blur appears in the warning line target searching method, the blur motionrecovery method is also proposed. The experimental results show that the activesearching increases target searching range, and effectively improve the probability oftarget successful acquisition.
To sum up, the innovations of the paper lie in: through the researching andimproving of existing techniques in the range of photoelectric tracking measurementsystem, the related key technologies of multi-feature information fusion and stabletracking is put forward when working at the background of clouds. The significance of the technology is to solve the problems of easily tracking interruption under thebackground of the clouds, to make up for the defect of current range target trackingtechnology which relyes too much on the weather. The method has certain practicalvalue in the field of target tracking measurement.
本文以靶场光电跟踪测量系统为研究平台，以提高云层背景下目标跟踪准确性和平稳性为目的，围绕当前云层背景下目标跟踪方法存在的不足之处，开展目标识别、目标跟踪及目标搜索相关关键技术的研究。在方法具有实用价值的指导思想下，提出靶场光电跟踪测量系统在云层背景下稳定跟踪的实现方案。全文完成的主要工作归纳如下：
研究了云层背景下疑似目标区域分割技术。针对现有的单一图像分割技术难以实现云层背景下疑似目标区域的有效分割，确立了依据跟踪模式特点采用相适应的图像分割技术思路。将目标跟踪过程划分成稳定跟踪模式和主动搜索模式。在稳定跟踪模式下，依据目标大致位置区域已知的特点，采用改进后的最大类间方差阈值分割方法进行快速图像分割；在主动搜索模式下，依据全视场分割的要求，提出了基于视觉注意机制的图像分割方法实现精准的疑似目标区域分割，通过算例验证了方法的可行性。
研究了目标多特征信息融合识别技术。针对目标周边存在云层或被云层遮挡时，采用单一的灰度特征易将云层错误识别成目标的问题，提出了基于神经网络和证据理论两级联合的多特征信息融合目标识别方法。首先提取分割后疑似目标区域的多种特征信息，对每种特征信息通过相应神经网络进行特征识别，再利用证据理论对多个神经网络的识别结果融合判决。该方法一方面利用了目标多特征信息降低目标误识别的概率，另一方面两级联合的信息融合模型体现出良好的可扩展性，能方便地根据不同情形、不同对象增减改变特征融合信息源，具有很强实用价值。实验结果证明该方法能准确辨别出云层和目标。
研究了目标运动状态估计反馈跟踪技术。针对传统的目标脱靶量直接反馈跟踪方式难以适应云层背景下目标跟踪需求的问题，深入研究了采用目标运动状态估计反馈的跟踪方法及实现细节。分析了不同应用情形下目标运动状态估计的可观测性问题，探讨了卡尔曼滤波、不敏卡尔曼滤波及交互多模型滤波三种滤波估计方法的应用改进，给出了实际应用时目标运动模型建立、数据时空对齐、数据关联及状态估计等实现过程。实验结果表明该方法弥补了直接反馈跟踪方式的跟踪性能严重依赖于测量数据质量的缺陷，提高了云层背景下目标跟踪稳定性。
研究了运动目标主动搜索策略。针对目标跟踪过程中受到云层遮挡而短暂丢失目标的情形，提出了基于最大概率区间的目标主动搜索方法；针对长时间丢失目标且缺乏外界引导数据的情形，提出了在拦截点处采用警戒线目标搜索方法来提高搜索效果，同时针对警戒线搜索过程中目标成像会出现运动模糊的情况，给出了运动模糊恢复方法。实验结果表明主动搜索增大了目标有效搜索区间，可以大幅提高目标搜索成功概率。
综上，本课题通过对靶场光电跟踪测量系统现有的目标跟踪技术研究和改进，提出并验证了在云层背景下工作时目标多特征信息融合及稳定跟踪的相关关键技术，该技术的意义在于解决了云层背景下目标跟踪易中断的问题，弥补了当前靶场光电跟踪测量系统的目标跟踪技术对天气条件过于依赖的不足，在目标跟踪测量领域具有一定的实用价值。
A range photoelectronic tracking measurement system is the use of opticalimaging techniques tracking measurement space flight target, in order to get highprecision measurement data of the target, and provide important basis for testevaluation and fault analysis of weapons and spaceflight. In recent years, the needsof new test mission and range construction development put forward higherrequirements for range photoelectric tracking measurement system to continuouslyand stably track target. However, large numbers of field test results show that thecurrent target tracking technology of electro-optical tracking measurement systemexists defects that it relied too much on the weather conditions, especially theinclusion of many clouds in the clear sky. On the one hand, it is easy to misidentifyclouds into the target which will leads to tracking interruption; on the other hand, itlacks the ability of effective searching and tracking again when the target getcovered, most of the time it takes a long time and rely on external guide data andpersonal judgment operation to search and track target again. The clear sky withclouds is a common meteorological condition in shooting range test task. To avoidmissing out on the follow-up measurement data and even the key period of valuabledata for the influence of partial clouds, it is necessary to study and solve the problemof the easily interruption with tracking under the background of clouds.
The paper is based on the range photoelectric tracking measurement system asthe research platform, on the purpose of improving the target tracking accuracy andstability as well as around the deficiency of current target tracking method under thebackground of the cloud, to carry out the relevant key technologies research of targetidentification, target tracking and target search. With the ideology of practicality, thestable tracking scheme is proposed under the background of clouds.
The main works accomplished in the paper summarized as follows:
The suspected target region segmentation technology is researched under thebackground of the clouds. Image segmentation can effectively reduce the operationcosts of subsequent target identification. In view of the existing single imagesegmentation technique is difficult to realize effective segmentation suspected targetarea under the background of clouds, an idea is established that image segmentationtechnology is adopted flexibly according to the characteristics of the tracking mode.The target tracking of range photoelectric tracking measurement system is composedwith stable tracking mode and active searching mode. In stable tracking mode,beacause the rough target location is known, the improved maxinum between-clustervariance threshold segmentation is adopted for fast image segmentation. In activesearching mode, due to the requirements of the full view-field segmentation, theimage segmentation method based on visual attention mechanism is proposed torealize precise suspected target area segmentation. The experimental results showthat the method is feasible.
The target multi-feature information fusion recognition technology is researched.If the target is surrounded or covered by clouds, the cloud is easily identified intotarget through the single gray characteristics. Thus, the multi-feature informationfusion recognition method is proposed basing on two-level jointing neutral networksand evidence theory. Various feature information of the segmented suspecting targetarea is extracted firstly, and then the corresponding neutral network is used to featurerecognize. Afterwards, evidence theory is utilized to incorporate multiple neuralnetwork outputs and make the synthetical decisions. On the one hand, the target multi-feature information is used to improve recognition accuracy. On the other hand,the two-level joint information fusion model shows well extensible. It has stronglypractical value that it can easily increase or decrease feature fusion informationsource according to different situations or different object. The experimental resultsshow the clouds and the target can be accurately identified through the method.
The target motion state estimation feedback tracking technology is researched.Because the traditional way of direct feedback tracking target is difficult to meet therequirement of target tracking under clouds background, the target motion stateestimation feedback tracking technology is studyed deeply in the range photoelectrictracking measurement system. The observability problem is analyzed in differentapplication situations. Kalman filter, unscented kalman filter and interactive multiplemodel filter are discussed and improved for the application. The target motion model,data space and time alignment, data association and state estimation are achieved.The experimental results show that this method makes up for defects that the targettracking performance of direct feedback seriously dependent on the measuring dataquality, and improves the stability of target tracking in the cloud backgrounds.
The active searching strategy of moving target is studied. It puts forward the targetactive searching method based on the maximum probability areas in the situation ofcloud-covering and temporary loss of target. In order to speed up the searchingresults for long losting goals and lack of external guide data, it puts forward thewarning line target searching method at the intercepting point. Because the imagemotion blur appears in the warning line target searching method, the blur motionrecovery method is also proposed. The experimental results show that the activesearching increases target searching range, and effectively improve the probability oftarget successful acquisition.
To sum up, the innovations of the paper lie in: through the researching andimproving of existing techniques in the range of photoelectric tracking measurementsystem, the related key technologies of multi-feature information fusion and stabletracking is put forward when working at the background of clouds. The significance of the technology is to solve the problems of easily tracking interruption under thebackground of the clouds, to make up for the defect of current range target trackingtechnology which relyes too much on the weather. The method has certain practicalvalue in the field of target tracking measurement.
引文
[1]于志坚．航天测控系统工程[M]．北京：国防工业出版社，2008．
[2]李尊民.电视图像自动跟踪的基本原理[M]．北京：国防工业出版社，1998．
[3]葛澎.一种快速边缘跟踪与提取的新算法[J].微电子学与计算机,2005,22(8):14-17.
[4]任仙怡,廖云涛,张桂林等.一种新的相关跟踪方法研究[J].中国图象图形学报,2002,6:35-39.
[5]沈永良,徐亚飞.舰载平台下的光电跟踪技术[J].火力与指挥控制,2008,7:13-15.
[6] M.Nelson, K.Mason. A model-based approach to information fusion[C].Proceedings of Information, Decision and Control,1999:395-400.
[7]李斌，陈以，韩元杰.模糊证据理论综述[J].兵工自动化,2005,24(3):79-81.
[8]王菲.一种被动声纳目标识别数据融合方法的研究[J].中国水运(理论版),2007,6(4):15-21.
[9]王洪锋，周磊，单甘霖.国外军事信息融合理论与应用的研究进展[J].电光与控制，2007,14(4):13-18.
[10] D. Y. Yang. Multi-sensor data fusion and its application to industrial control[C].Proceedings of the39th SICE Annual Conference,2000:215-220.
[11]张红蕾，宋建社.信息融合技术在SAR图像边缘检测中的应用[J].计算机工程与应用,2003,27(1):15-17.
[12] D. L. Hall, J. Llinas. An introduction to multisensor data fusion[J]. Proceedingsof the IEEE,1997,85(1):6-23.
[13]张池平.多传感器信息融合方法及其在空间目标识别中的应用[D]:[博士学位论文].哈尔滨:哈尔滨工业大学航天学院,2006.
[14] K.Yanagi, S. Furuichi, K. Kuriyama. A generalized skew information anduncertainty relation[J]. IEEE Information Theory Society,2005,51(12):4401-4404.
[15]朱泽君,黄涛,刘曦霞等.多传感器数据融合技术研究现状及发展方向[J].舰船电子工程,2009,29(2):13-16.
[16]陈艳琴,罗大庸.多信息融合技术及在无损检测中的应用[J].电子科技大学学报,2003,32(6):661-664.
[17]王改革．基于智能算法的目标威胁估计[D]:[博士学位论文].长春：中国科学院研究生院（长春光学精密机械与物理研究所）机械电子工程，2013.
[18] A. D. Bimbo, F. Dini. Particle filter-based visual tracking with a first orderdynamic model and uncertainty adaptation[J]. Computer Vision and ImageUnderstanding,2011,6:15-20.
[19]胡文刚.基于多值域特征及数据融合的焊缝缺陷超声检测与识别[D]:[博士学位论文].哈尔滨：哈尔滨工业大学材料科学与工程学院，2012.
[20]周玉龙,何永强,张维安.红外鱼眼系统下的多目标威胁评估研究[J].光学学报,2012,6:6110011-6110017.
[21] Y. Bar-Shalom, X. R. Li. Estimation and tracking: principles, techniques, andsoftware[M]. Boston: Artech House,1993.
[22] R. E. Kalman. A New Approach to Linear Filtering and Prediction Problems[J].Journal of Basic Engineering,1960:35-45.
[23] R. E. Kalman, R. S. Bucy. New results in linear filtering and prediction theory[J].Journal of Basic Engineering,1961:95–108.
[24] F. Gustafsson, G. Hendeby. Some Relations Between Extended and UnscentedKalman Filters[J]. IEEE Transactions on Signal Processing,2012,60(2):545-555.
[25] G. A. Einicke, L. B. White. Robust Extended Kalman Filtering[J]. IEEETransactions Signal Processing,1999,47(9):2596-2599.
[26] V. J. Aidala, S. C. Nardone. Biased Estimation Properties of the Pseudo LinearTracking Filter[J]. IEEE Transactions on Aerospace Electronics Systems,1982,18(4):432-441.
[27] S. K. Rao. Pseudo–linear Estimator for Bearings-only Passive Target Tracking[J].IEE Proceedings of Radar Sonar and Navigation,2001,148(1):16-22.
[28] S. Julier, J. Uhlmann. A new extension of the kalman filter to nonlinearsystems[C].11thInt Symposium Aerospace/Defense Sensing, Simulation and Controls,Orlando,1997:54-65.
[29] S. Julier, J. Uhlmann, H. F. Durrant-Whyte. A new method for the nonlineartransformation of means and covariances in filters and estimators[J]. IEEETransactions on Automatic Control,2000,45(3):477-482.
[30] A. Doueet, N. Gordon, V. KTishnamurthy. Partical filters for state estimation onjump markov linear systems[J]. IEEE Transactions on Signal Proeessing,2001,49(3):613~624.
[31] M. R. Morelande, S. Challa. Maneuvering target tracking in clutter using particalfilters[J]. IEEE Transactions on Aerospace and Electronic Systems,2005,41(l):252-270.
[32] R. A. Singer, J. J. Stein. An optimal tracking filter for processing sensor data ofimprecisely determined origin in surveillance systems[C]. Proceedings of the1971IEEE Conference on Decision and Control,1971:171-175.
[33] Y. Bar-Shalom, A. G. Jaffer. Adaptive nonlinear filtering for tracking withmeasurement of uncertain origin[C]. Proceedings of the1972IEEE Conference onDecision and Control,1972:243-247.
[34] Y. Bar-Shalom, E. Tse. Tracking in a cluttered environment with probabilisticdata association[J]. Automatica,1975,11(5):451-460.
[35] D. B. Reid. An algorithm for tracking multiple targets[J]. IEEE Transactions onAutomatic Control,1979,24(6):843-854.
[36] R. A. Singer. Estimating optimal tracking filter performance for mannedmaneuvering targets[J]. IEEE Transactions on AES,1970,6(4):473-483.
[37]周宏仁,敬忠良,王培德.机动目标跟踪[M].北京：国防工业出版社，1991.
[38] K. Melirotra, P. R. Mahapatra. A jerk model to tracking highly maneuveringtargets[J]. IEEE Transactions on AES,1997,33(4):1094-1105.
[39] H. A. Blom, Y. Bar-Shalom, J. Dayan. The interacting multiple model algorithmfor systems with markovian switching coefficients[C]. IEEE Transactions onAutomatic Control,1988,33(8):780-783.
[40] A. Munir, D. P. Atherton. Maneuvering target tracking using an adaptiveinteracting multiple model algorithm[C]. Proeeedings of the American ControlConference Baltimore,1994:1324-1328.
[41] M. Sonka, V. Hlavac, R.Boyle. Image processing, analysis, and machinevision[M]. Beijing: Posts and Telecom Press,2003.
[42] R. C. Gonzalez, R. E. Woods著，阮秋琦，阮宇智等译．数字图像处理(第二版)[M]．北京：电子工业出版社，2003．
[43] N. Otsu. A threshold selection method from gray-level histograms[J]. IEEETransactions on System, Man and Cybernetics,1979,9(1):62-66.
[44] A. Nakib, H. Oulhadj, P. Siarry. Fractional differentiation and non-Paretomultiobjective optimization for image thresholding[J]. Engineering Applications ofArtificial Intelligence,2009,22:236-249.
[45] H. F. Ng. Automatic thresholding for defect detection[J]. Pattern RecognitionLetters,2006,27(14):1644-1649.
[46] Y. Wang, T. Adaii, S. Y. Kung. Quantification and segmentation of brain tissuesfrom MR images a probabilistic neural networks approach[J]. IEEE Transactions onImage Processing,1998,7(8):1-12.
[47] G. Kuntimad, H. S. Ranganath. Perfect image segmentation using pulse coupledneural networks[J]. IEEE Transactions on Neural Networks,1999,10(3):591-598.
[48] J. Y. Zhang, X. J. Fan, J. Y. Dong, et al. Image segmentation based on modifiedpulse-coupled neural networks[J]. Chinese Journal of Electronics,2007,16(1):119-122.
[49] H. Berg, R. Olsson, T. Lindblad, et al. Automatic design of pulse coupled neuronsfor image segmentation[J]. Neurocomputing,2008,71(6):1980-1993.
[50] O. Chapelle, P. Haffner, V. N. Vapnik. Support vector machines for histogram-based image classification [J]. IEEE Transactions on Neural Networks,1999,10(5):1055-1064.
[51]徐海祥,喻莉,朱光喜等.基于支持向量机的磁共振脑组织图像分割[J].中国图象图形学报,2005,10(10):1275-1280.
[52]魏鸿磊,欧宗瑛,张建新.采用支持向量机的指纹图像分割[J].系统仿真学报,2007,19(5):2362-2365.
[53] C. Koch, S. Ullman. Shifts in selective visual attention: towards the underlyingneural circuitry[J]. Human neurobiology,1985,4(4):219.
[54] L. Itti, C. Koch, E. Niebur. A model of saliency-based visual attention for rapidscene analysis[J]. IEEE Transaction on PAMI,1998,20(11):1254-1259.
[55] L. Itti, C. Koch. Feature combination strategies for saliency-based visualattentions systems[J]. Journal of Electronic Imaging,2001,10(1):161-169.
[56] A. Treisman, G. Gelade. A feature integration theory of attention[J]. CognitivePsychology,1980,12:97-136.
[57] D. G. Lowe. Distinctive image features from scale-invariant keypoints[J].International journal of computer vision,2004,60(2):91-110.
[58]周全,张荣,尹东.基于高斯金字塔的遥感云图多尺度特征提取[J].遥感技术与应用,2010,25(5):604-608.
[59] T. S. Lee. Image representation using2D Gabor wavelets[J]. IEEE Transactionson Pattern Analysis and Machine Intelligence,1996,18(10):959-971.
[60] J. G. Daugman. Two Dimensional Spectral Analysis of Cortical Receptive FieldProfiles. Vision Research,1980,20:847-856.
[61]封子军,张晓玲,张靖波.各向异性扩散与高斯差分模型的多聚焦图像融合方法[J].中国图象图形学报,2011,11:2054-2059.
[62]陶唐飞,韩崇昭,代雪峰等.综合边缘检测和区域生长的红外图像分割方法[J],光电工程,2004,10:50-52
[63] A. S. Pandya, R. B. Macy著，徐勇，荆涛等译．神经网络模式识别及实现[M]．北京：电子工业出版社，1999．
[64]边肇祺，张学工等．模式识别[M]．北京：清华大学出版社，1999．
[65] F. D. Frate, F. Pacifici, G. Schiavon, et al. Use of neural networks for automaticclassification from high-resolution images[J]. IEEE Transactions on Geosci RemoteSens,2007,45(4):800-809.
[66] D. S Zhang, G. J. Lu. Shape-based image retrieval using generic Fourierdescriptor[J]. Signal Processing: Image Communication,2002,17:825-848.
[67] S. Kuthirummal, C. V. Jawahar, P. J. Narayanan. Fourier domain representationof planar curves for recognition in multiple views[J]. Pattern Recognition,2004,37:739-754.
[68]李利，范九伦.基于Radon变换的指纹图像滤波增强方法[J].中国图象图形学报,2009,3:417-421.
[69]李军宏,潘泉,陈玉春等.基于Radon变换的矩不变量及其应用[J].西北工业大学学报,2004,3:392-396.
[70]张锁平,张春田.基于Radon变换和梯度跟踪的海岸线提取算法[J].光电工程,2009,36(3):57-62.
[71]王继成，蔡义发，吕维雪．一种自动生成神经网络结构的新方法[J]．自动化学报，1996，22(01)：19-25.
[72]许进，焦李成，保铮．神经网络与图论[J]．系统工程与电子技术，1995(05)：31-43.
[73]何述东，瞿坦，黄献青，黄心汉．多层前向神经网络结构的研究进展[J]．控制理论与应用，1998，15(03)：313-319.
[74] D. E. Rumelhart, J. L. McClelland. Parallel Distributed Processing: Explorationsin the Microstructure of Cognition[M]. USA: The MIT press,1986:1-567.
[75] G. Shafer. A Mathematical Theory of Evidence[M]. USA: Princeton universitypress,1976:1-302.
[76] A. Dempster. Upper and Lower Probabilities Induced by a MultiplicandMapping[J]. Annals of Mathematical Statistics,1967,38:325-339.
[77] J. Daniel, J. P. Lauffenburger, M. Basset. Multi-Criterion Dempster-ShaferFusion for Speed Limit Determination[C]: Preprints of the18th IFAC WorldCongress,2011:668-673.
[78]肖人彬，费奇．基于证据理论的不确定性分析[J]．华中理工大学学报，1993，21(03)：24-30.
[79] A. Dempster. A Generalization of Bayesian Inference[J]. Journal of the RoyalStatistical Society,1968,30:205-245.
[80]孙全，叶秀清，顾伟康．一种新的基于证据理论的合成公式[J]．电子学报，2000(08)：117-119.
[81]向阳，史习智．证据理论合成规则的一点修正[J]．上海交通大学学报，1999，33(03)：357-360.
[82] C. Jauffret, D. Pillon. Observability in Passive Target Motion Analysis[J]. IEEETransactions on Aerospace and Electronic System,1996,32(4):1290-1300.
[83] V. J. Aidala. Kalman filter behavior in bearing-only tracking application. IEEETransactions on Aerospace and Electronic System,1979,15(1):29-39.
[84]辛云宏，杨万海．基于伪线性卡尔曼滤波的两站红外无源定位及跟踪技术[J].西安电子科技大学学报(自然科学版),2004,31(4):505-508.
[85]徐肖豪，高彦杰，杨国庆．交互式多模型算法中模型集选择的分析研究[J]．航空学报，2004，25(4)：352-356．
[86]刘建书，李人厚，张贞耀等．交互式多模型算法的模型集设计[J]．控制与决策，2007，22(3)：326-332．
[87] J. N. Sanders-Reed. Multi-Target, Multi-Sensor, Closed Loop Tracking[J].Proceedings of SPIE,2004,5430:94-112.
[88] R. E. Helmick, T. R. Rice. Removal of alignment errors in an integrated systemof two3D sensors[J]. IEEE Transactions on Aerospace and Electronic Systems,1993,29(4):1333-1343.
[89] N. Nabaa, R. H. Bishop. Solution to a multisensor tracking problem with sensorregistration errors[J]. IEEE Transactions on Aerospace and Electronic Systems,1999,35(1):354-363.
[90] D. C. Cowley, B. Shafai. Registration in multi-sensor data fusion and tracking[C].Proceedings of the American Control Conference,1993,875–879.
[91] T. T. Fjallbrant. Interpolation and extrapolation in nonuniform samplingsequences with average sampling rates below the Nyquist rate[J]. Electronics Letters,1975,11(12):264-266.
[92]刘蛟龙,尉建利,于云峰.弹道导弹自由段拦截窗口建模与仿真[J].指挥控制与仿真,2012,06:72-76.
[93]李长生．军事运筹学教程[M]．北京：军事科学出版社，2000．
[94]涂娇姣，洪汉玉，张天序．保边缘的递归逆滤波盲目图像复原算法[J]．计算机与数字工程，2006，34(12)：4-6．
[95]陈湘凭，杨大磊，陈玺．图像的运动模糊复原技术研究[J]．计算机与数字工程，2008，36(8)：136-139．
[96] L. Yuan, J. Sun, L. Quan, et al. Progressive Inter-Scale and Intra-Scale Non-BlindImage Deconvolution[J]. ACM Transactions on Graphics,2008,27(3):1-10.
[97]范海菊，刘艳芳，冯乃勤．运动模糊图像点扩散函数参数识别的新方法[J]．计算机工程与应用，2011，47(30)：188-191．
[98]赵琳，金伟其，陈翼男等．一种新的运动模糊图像超分辨力盲复原算法[J]．光子学报，2008，37(11)：2355-2359．
[99]王书振，邹子健，李莉等．一种基于稀疏正则化的图像盲复原方法[J]．西安电子科技大学学报(自然科学版)，2012，39(6)：167-169．
[100]吴显金，王润生．基于边缘恢复和伪像消除的正则化图像复原[J]．电子与信息学报，2006，28(4)：577-581．
[101]石明珠，许廷发，张坤．运动成像混合模糊的全变分图像复原[J]．光学精密工程，2011，19(8)：1973-1981．
[2]李尊民.电视图像自动跟踪的基本原理[M]．北京：国防工业出版社，1998．
[3]葛澎.一种快速边缘跟踪与提取的新算法[J].微电子学与计算机,2005,22(8):14-17.
[4]任仙怡,廖云涛,张桂林等.一种新的相关跟踪方法研究[J].中国图象图形学报,2002,6:35-39.
[5]沈永良,徐亚飞.舰载平台下的光电跟踪技术[J].火力与指挥控制,2008,7:13-15.
[6] M.Nelson, K.Mason. A model-based approach to information fusion[C].Proceedings of Information, Decision and Control,1999:395-400.
[7]李斌，陈以，韩元杰.模糊证据理论综述[J].兵工自动化,2005,24(3):79-81.
[8]王菲.一种被动声纳目标识别数据融合方法的研究[J].中国水运(理论版),2007,6(4):15-21.
[9]王洪锋，周磊，单甘霖.国外军事信息融合理论与应用的研究进展[J].电光与控制，2007,14(4):13-18.
[10] D. Y. Yang. Multi-sensor data fusion and its application to industrial control[C].Proceedings of the39th SICE Annual Conference,2000:215-220.
[11]张红蕾，宋建社.信息融合技术在SAR图像边缘检测中的应用[J].计算机工程与应用,2003,27(1):15-17.
[12] D. L. Hall, J. Llinas. An introduction to multisensor data fusion[J]. Proceedingsof the IEEE,1997,85(1):6-23.
[13]张池平.多传感器信息融合方法及其在空间目标识别中的应用[D]:[博士学位论文].哈尔滨:哈尔滨工业大学航天学院,2006.
[14] K.Yanagi, S. Furuichi, K. Kuriyama. A generalized skew information anduncertainty relation[J]. IEEE Information Theory Society,2005,51(12):4401-4404.
[15]朱泽君,黄涛,刘曦霞等.多传感器数据融合技术研究现状及发展方向[J].舰船电子工程,2009,29(2):13-16.
[16]陈艳琴,罗大庸.多信息融合技术及在无损检测中的应用[J].电子科技大学学报,2003,32(6):661-664.
[17]王改革．基于智能算法的目标威胁估计[D]:[博士学位论文].长春：中国科学院研究生院（长春光学精密机械与物理研究所）机械电子工程，2013.
[18] A. D. Bimbo, F. Dini. Particle filter-based visual tracking with a first orderdynamic model and uncertainty adaptation[J]. Computer Vision and ImageUnderstanding,2011,6:15-20.
[19]胡文刚.基于多值域特征及数据融合的焊缝缺陷超声检测与识别[D]:[博士学位论文].哈尔滨：哈尔滨工业大学材料科学与工程学院，2012.
[20]周玉龙,何永强,张维安.红外鱼眼系统下的多目标威胁评估研究[J].光学学报,2012,6:6110011-6110017.
[21] Y. Bar-Shalom, X. R. Li. Estimation and tracking: principles, techniques, andsoftware[M]. Boston: Artech House,1993.
[22] R. E. Kalman. A New Approach to Linear Filtering and Prediction Problems[J].Journal of Basic Engineering,1960:35-45.
[23] R. E. Kalman, R. S. Bucy. New results in linear filtering and prediction theory[J].Journal of Basic Engineering,1961:95–108.
[24] F. Gustafsson, G. Hendeby. Some Relations Between Extended and UnscentedKalman Filters[J]. IEEE Transactions on Signal Processing,2012,60(2):545-555.
[25] G. A. Einicke, L. B. White. Robust Extended Kalman Filtering[J]. IEEETransactions Signal Processing,1999,47(9):2596-2599.
[26] V. J. Aidala, S. C. Nardone. Biased Estimation Properties of the Pseudo LinearTracking Filter[J]. IEEE Transactions on Aerospace Electronics Systems,1982,18(4):432-441.
[27] S. K. Rao. Pseudo–linear Estimator for Bearings-only Passive Target Tracking[J].IEE Proceedings of Radar Sonar and Navigation,2001,148(1):16-22.
[28] S. Julier, J. Uhlmann. A new extension of the kalman filter to nonlinearsystems[C].11thInt Symposium Aerospace/Defense Sensing, Simulation and Controls,Orlando,1997:54-65.
[29] S. Julier, J. Uhlmann, H. F. Durrant-Whyte. A new method for the nonlineartransformation of means and covariances in filters and estimators[J]. IEEETransactions on Automatic Control,2000,45(3):477-482.
[30] A. Doueet, N. Gordon, V. KTishnamurthy. Partical filters for state estimation onjump markov linear systems[J]. IEEE Transactions on Signal Proeessing,2001,49(3):613~624.
[31] M. R. Morelande, S. Challa. Maneuvering target tracking in clutter using particalfilters[J]. IEEE Transactions on Aerospace and Electronic Systems,2005,41(l):252-270.
[32] R. A. Singer, J. J. Stein. An optimal tracking filter for processing sensor data ofimprecisely determined origin in surveillance systems[C]. Proceedings of the1971IEEE Conference on Decision and Control,1971:171-175.
[33] Y. Bar-Shalom, A. G. Jaffer. Adaptive nonlinear filtering for tracking withmeasurement of uncertain origin[C]. Proceedings of the1972IEEE Conference onDecision and Control,1972:243-247.
[34] Y. Bar-Shalom, E. Tse. Tracking in a cluttered environment with probabilisticdata association[J]. Automatica,1975,11(5):451-460.
[35] D. B. Reid. An algorithm for tracking multiple targets[J]. IEEE Transactions onAutomatic Control,1979,24(6):843-854.
[36] R. A. Singer. Estimating optimal tracking filter performance for mannedmaneuvering targets[J]. IEEE Transactions on AES,1970,6(4):473-483.
[37]周宏仁,敬忠良,王培德.机动目标跟踪[M].北京：国防工业出版社，1991.
[38] K. Melirotra, P. R. Mahapatra. A jerk model to tracking highly maneuveringtargets[J]. IEEE Transactions on AES,1997,33(4):1094-1105.
[39] H. A. Blom, Y. Bar-Shalom, J. Dayan. The interacting multiple model algorithmfor systems with markovian switching coefficients[C]. IEEE Transactions onAutomatic Control,1988,33(8):780-783.
[40] A. Munir, D. P. Atherton. Maneuvering target tracking using an adaptiveinteracting multiple model algorithm[C]. Proeeedings of the American ControlConference Baltimore,1994:1324-1328.
[41] M. Sonka, V. Hlavac, R.Boyle. Image processing, analysis, and machinevision[M]. Beijing: Posts and Telecom Press,2003.
[42] R. C. Gonzalez, R. E. Woods著，阮秋琦，阮宇智等译．数字图像处理(第二版)[M]．北京：电子工业出版社，2003．
[43] N. Otsu. A threshold selection method from gray-level histograms[J]. IEEETransactions on System, Man and Cybernetics,1979,9(1):62-66.
[44] A. Nakib, H. Oulhadj, P. Siarry. Fractional differentiation and non-Paretomultiobjective optimization for image thresholding[J]. Engineering Applications ofArtificial Intelligence,2009,22:236-249.
[45] H. F. Ng. Automatic thresholding for defect detection[J]. Pattern RecognitionLetters,2006,27(14):1644-1649.
[46] Y. Wang, T. Adaii, S. Y. Kung. Quantification and segmentation of brain tissuesfrom MR images a probabilistic neural networks approach[J]. IEEE Transactions onImage Processing,1998,7(8):1-12.
[47] G. Kuntimad, H. S. Ranganath. Perfect image segmentation using pulse coupledneural networks[J]. IEEE Transactions on Neural Networks,1999,10(3):591-598.
[48] J. Y. Zhang, X. J. Fan, J. Y. Dong, et al. Image segmentation based on modifiedpulse-coupled neural networks[J]. Chinese Journal of Electronics,2007,16(1):119-122.
[49] H. Berg, R. Olsson, T. Lindblad, et al. Automatic design of pulse coupled neuronsfor image segmentation[J]. Neurocomputing,2008,71(6):1980-1993.
[50] O. Chapelle, P. Haffner, V. N. Vapnik. Support vector machines for histogram-based image classification [J]. IEEE Transactions on Neural Networks,1999,10(5):1055-1064.
[51]徐海祥,喻莉,朱光喜等.基于支持向量机的磁共振脑组织图像分割[J].中国图象图形学报,2005,10(10):1275-1280.
[52]魏鸿磊,欧宗瑛,张建新.采用支持向量机的指纹图像分割[J].系统仿真学报,2007,19(5):2362-2365.
[53] C. Koch, S. Ullman. Shifts in selective visual attention: towards the underlyingneural circuitry[J]. Human neurobiology,1985,4(4):219.
[54] L. Itti, C. Koch, E. Niebur. A model of saliency-based visual attention for rapidscene analysis[J]. IEEE Transaction on PAMI,1998,20(11):1254-1259.
[55] L. Itti, C. Koch. Feature combination strategies for saliency-based visualattentions systems[J]. Journal of Electronic Imaging,2001,10(1):161-169.
[56] A. Treisman, G. Gelade. A feature integration theory of attention[J]. CognitivePsychology,1980,12:97-136.
[57] D. G. Lowe. Distinctive image features from scale-invariant keypoints[J].International journal of computer vision,2004,60(2):91-110.
[58]周全,张荣,尹东.基于高斯金字塔的遥感云图多尺度特征提取[J].遥感技术与应用,2010,25(5):604-608.
[59] T. S. Lee. Image representation using2D Gabor wavelets[J]. IEEE Transactionson Pattern Analysis and Machine Intelligence,1996,18(10):959-971.
[60] J. G. Daugman. Two Dimensional Spectral Analysis of Cortical Receptive FieldProfiles. Vision Research,1980,20:847-856.
[61]封子军,张晓玲,张靖波.各向异性扩散与高斯差分模型的多聚焦图像融合方法[J].中国图象图形学报,2011,11:2054-2059.
[62]陶唐飞,韩崇昭,代雪峰等.综合边缘检测和区域生长的红外图像分割方法[J],光电工程,2004,10:50-52
[63] A. S. Pandya, R. B. Macy著，徐勇，荆涛等译．神经网络模式识别及实现[M]．北京：电子工业出版社，1999．
[64]边肇祺，张学工等．模式识别[M]．北京：清华大学出版社，1999．
[65] F. D. Frate, F. Pacifici, G. Schiavon, et al. Use of neural networks for automaticclassification from high-resolution images[J]. IEEE Transactions on Geosci RemoteSens,2007,45(4):800-809.
[66] D. S Zhang, G. J. Lu. Shape-based image retrieval using generic Fourierdescriptor[J]. Signal Processing: Image Communication,2002,17:825-848.
[67] S. Kuthirummal, C. V. Jawahar, P. J. Narayanan. Fourier domain representationof planar curves for recognition in multiple views[J]. Pattern Recognition,2004,37:739-754.
[68]李利，范九伦.基于Radon变换的指纹图像滤波增强方法[J].中国图象图形学报,2009,3:417-421.
[69]李军宏,潘泉,陈玉春等.基于Radon变换的矩不变量及其应用[J].西北工业大学学报,2004,3:392-396.
[70]张锁平,张春田.基于Radon变换和梯度跟踪的海岸线提取算法[J].光电工程,2009,36(3):57-62.
[71]王继成，蔡义发，吕维雪．一种自动生成神经网络结构的新方法[J]．自动化学报，1996，22(01)：19-25.
[72]许进，焦李成，保铮．神经网络与图论[J]．系统工程与电子技术，1995(05)：31-43.
[73]何述东，瞿坦，黄献青，黄心汉．多层前向神经网络结构的研究进展[J]．控制理论与应用，1998，15(03)：313-319.
[74] D. E. Rumelhart, J. L. McClelland. Parallel Distributed Processing: Explorationsin the Microstructure of Cognition[M]. USA: The MIT press,1986:1-567.
[75] G. Shafer. A Mathematical Theory of Evidence[M]. USA: Princeton universitypress,1976:1-302.
[76] A. Dempster. Upper and Lower Probabilities Induced by a MultiplicandMapping[J]. Annals of Mathematical Statistics,1967,38:325-339.
[77] J. Daniel, J. P. Lauffenburger, M. Basset. Multi-Criterion Dempster-ShaferFusion for Speed Limit Determination[C]: Preprints of the18th IFAC WorldCongress,2011:668-673.
[78]肖人彬，费奇．基于证据理论的不确定性分析[J]．华中理工大学学报，1993，21(03)：24-30.
[79] A. Dempster. A Generalization of Bayesian Inference[J]. Journal of the RoyalStatistical Society,1968,30:205-245.
[80]孙全，叶秀清，顾伟康．一种新的基于证据理论的合成公式[J]．电子学报，2000(08)：117-119.
[81]向阳，史习智．证据理论合成规则的一点修正[J]．上海交通大学学报，1999，33(03)：357-360.
[82] C. Jauffret, D. Pillon. Observability in Passive Target Motion Analysis[J]. IEEETransactions on Aerospace and Electronic System,1996,32(4):1290-1300.
[83] V. J. Aidala. Kalman filter behavior in bearing-only tracking application. IEEETransactions on Aerospace and Electronic System,1979,15(1):29-39.
[84]辛云宏，杨万海．基于伪线性卡尔曼滤波的两站红外无源定位及跟踪技术[J].西安电子科技大学学报(自然科学版),2004,31(4):505-508.
[85]徐肖豪，高彦杰，杨国庆．交互式多模型算法中模型集选择的分析研究[J]．航空学报，2004，25(4)：352-356．
[86]刘建书，李人厚，张贞耀等．交互式多模型算法的模型集设计[J]．控制与决策，2007，22(3)：326-332．
[87] J. N. Sanders-Reed. Multi-Target, Multi-Sensor, Closed Loop Tracking[J].Proceedings of SPIE,2004,5430:94-112.
[88] R. E. Helmick, T. R. Rice. Removal of alignment errors in an integrated systemof two3D sensors[J]. IEEE Transactions on Aerospace and Electronic Systems,1993,29(4):1333-1343.
[89] N. Nabaa, R. H. Bishop. Solution to a multisensor tracking problem with sensorregistration errors[J]. IEEE Transactions on Aerospace and Electronic Systems,1999,35(1):354-363.
[90] D. C. Cowley, B. Shafai. Registration in multi-sensor data fusion and tracking[C].Proceedings of the American Control Conference,1993,875–879.
[91] T. T. Fjallbrant. Interpolation and extrapolation in nonuniform samplingsequences with average sampling rates below the Nyquist rate[J]. Electronics Letters,1975,11(12):264-266.
[92]刘蛟龙,尉建利,于云峰.弹道导弹自由段拦截窗口建模与仿真[J].指挥控制与仿真,2012,06:72-76.
[93]李长生．军事运筹学教程[M]．北京：军事科学出版社，2000．
[94]涂娇姣，洪汉玉，张天序．保边缘的递归逆滤波盲目图像复原算法[J]．计算机与数字工程，2006，34(12)：4-6．
[95]陈湘凭，杨大磊，陈玺．图像的运动模糊复原技术研究[J]．计算机与数字工程，2008，36(8)：136-139．
[96] L. Yuan, J. Sun, L. Quan, et al. Progressive Inter-Scale and Intra-Scale Non-BlindImage Deconvolution[J]. ACM Transactions on Graphics,2008,27(3):1-10.
[97]范海菊，刘艳芳，冯乃勤．运动模糊图像点扩散函数参数识别的新方法[J]．计算机工程与应用，2011，47(30)：188-191．
[98]赵琳，金伟其，陈翼男等．一种新的运动模糊图像超分辨力盲复原算法[J]．光子学报，2008，37(11)：2355-2359．
[99]王书振，邹子健，李莉等．一种基于稀疏正则化的图像盲复原方法[J]．西安电子科技大学学报(自然科学版)，2012，39(6)：167-169．
[100]吴显金，王润生．基于边缘恢复和伪像消除的正则化图像复原[J]．电子与信息学报，2006，28(4)：577-581．
[101]石明珠，许廷发，张坤．运动成像混合模糊的全变分图像复原[J]．光学精密工程，2011，19(8)：1973-1981．