光电跟踪测量系统的动态成像分辨率分析及优化设计

光电跟踪测量系统的测量精度与成像质量受成像系统动态分辨率的影响.成像系统动态分辨率受光电跟踪测量系统跟踪精度的影响.分析与试验表明,光电跟踪测量系统的动态成像分辨率主要取决于静态分辨率和跟踪精度;工作过程中的光电跟踪测量系统的动态成像分辨率比光电跟踪测量系统的静态成像分辨率低;与传感器的积分时间相比,高频目标与低频目标对动态成像分辨率的影响模型不同;在传感器积分时间、目标的运动状态及对动态分辨率要求确定的条件下,可对静态成像分辨与控制系统进行优化设计.     

机载光电跟踪系统连续帧图像快速识别定位

针对机载光电跟踪系统连续帧图像快速识别、准确定位的要求,本文在深入分析连续图像序列特性和机载光电跟踪系统特点的基础上,结合精密转台的运动变化信息,提出了基于转角信息的连续帧图像处理算法.该方法根据精密转台的转角变化信息,采用连续的图像序列进行多帧间差值运算以获得目标残差图,残差图之间相乘.除去大部分随机噪声点,结合中值滤波和自适应渡门跟踪算法,大大的减小了运算量,实现了运动目标的快速跟踪定位.通过对目标实测图像序列的试验,结果表明,该算法具有快速、稳定等优点,能满足机载光电跟踪系统实时图像跟踪的要求.     

对光电成像系统干扰效果的评估方法

提出了对两类光电成像系统干扰效果的评估准则。对观瞄用光电成像系统,可以根据对成像功能的损伤破坏程度将激光致盲干扰效果划分为四个等级,根据对目标识别的影响程度将烟幕干扰效果划分为三个等级。为定量评估烟幕对观瞄用光电成像系统的干扰效果,可以采用成像系统在实施干扰前后图像之间的相关函数作为评估指标,按照相关函数值的大小将干扰效果划分为三个等级。对跟踪用光电成像系统,可依据有干扰时成像系统能否输出有效跟踪误差以及跟踪误差是否超出正常跟踪精度的三倍判定干扰是否有效,并依据干扰成功率的大小划分干扰等级。     

实时电视跟踪系统

应用于二维成像的实时图像识别,目前由于受到处理速度和信息带宽的严重限制,而只处于琐碎的或近琐碎的算法之中。目标识别以及用电视(视频)速率的图像识别在跟踪上的应用,是个普遍关心的问题。但是,过去的想法只限于非常简单的阈值或相关(限制窗口)方法。新的高速算法同快速数字硬件结合在一起产生了用于识别和跟踪导弹以及飞机的系统,该系统具有“智力”,这使其性能远远超过了以前设备的实时跟踪能力。基于分类算法的自适应统计群和投影已用于实时识别和跟踪明显变化的目标,这种变化是由于复杂的和不稳定的背景/前景所引起的。快速估值和预测算法都是将线性估值和平方估值相结合,以获得速度和灵敏度信息。在数据和结果的判定中,为了得到一个置信度量值,需要确定权。为了最大限度地保持跟踪,研究了几种方案。本文着重于该系统的理论方面论述,并且也讨论了用于得到实时跟踪方面的一些技术问题。     

基于ARM的嵌入式停车场数字图像监控系统

本文设计了基于ARM9 S3C2410的嵌入式图像采集处理系统，在内嵌工Linux操作系统的基础上实现了并行数据总线／USB接口图像接入、图像快速处理、IP化数据传输。根据停车场监控管理的需要，设计了一种基于连续图像帧差法的目标识别快速算法，完成了可疑目标的实时自动识别与跟踪、异常事件识别功能，并实现重要图像信息的本地压缩存储与IP化传输。     

一种新的红外与可见光图像融合与跟踪方法

多传感器图像融合技术在目标检测与跟踪领域中有广泛应用。提出了一种红外与可见光图像的新颖的特征级融合与运动目标跟踪方法。将目标轮廓用动态轮廓线表示,在目标跟踪过程中对于两类模式图像中的目标轮廓控制点向量,利用 B 样条形状空间模型将目标轮廓的特征级融合转换为控制点向量差的 L2 范数平方极小化。这种方法不需要图像配准,降低了融合的计算复杂度。同时,使用了自适应 Kalman 滤波技术,提高了动态轮廓线特征搜索的准确性。对比跟踪实验表明,融合后可见光图像的平均跟踪误差减小了 56.96%。     

多传感器模糊融合跟踪算法

针对集中式融合结构跟踪系统,利用随机逼近算法分析了权值的最优分配原则,提出了一种基于模糊推理的多传感器融合跟踪算法。该算法采用协方差匹配技术,依据滤波新息,动态调整测量噪声方差,使融合系统的均方误差始终最小。同时利用双滤波器结构,根据系统方差,实现滤波器间的动态切换,提出了基于模糊推理的并行双Unscented卡尔曼滤波自适应跟踪算法,增强当前统计模型对弱机动目标的适应能力。针对机动和非机动飞行航路进行了算法仿真,结果表明,在时变测量噪声条件下,采用模糊融合跟踪算法前后的速度均方根误差分别为45.7m/s和36.2m/s, 18.7m/s和9.6m/s,提高了多传感器系统的稳健性和跟踪精度。     

预测滤波技术在光电经纬仪中的应用仿真

针对光电经纬仪的电视跟踪系统脱靶量动态滞后问题,引入了预测滤波技术,分析指出经纬仪跟踪的目标在摄像机靶面上的成像点运动为机动,且认为机动加速度服从修正瑞利分布,采用基于机动目标“当前”统计模型的自适应卡尔曼滤波算法进行误差动态补偿,仿真结果表明,该算法较传统方法不但补偿精度高且极大抑制了随机误差。     

一种弱小目标的自适应搜索方法

针对成像跟踪系统中弱小目标搜索定位的难题,提出了一种新的自适应搜索算法,即利用全局寻优的自适应遗传算法搜索目标,以目标的多特征融合信息作为最佳定位的判断准则和适应度函数。实验结果表明,该方法不仅提高了检测精度,也改善了跟踪算法的稳定性和智能策略。     

实时动态场景生成

利用三维建模软件对目标和光电跟踪测量系统等实物建立三维模型,把捕获系统实时获取的目标状态信息和光电跟踪测量系统运行信息通过网络实时传送给场景生成计算机,场景生成计算机利用这些数据信息驱动三维模型,从而实时模拟目标与光电跟踪测量系统的运行状态。该系统可为指挥控制人员的决策提供更直观、简便的帮助。     

基于多机器人自组织协作的多目标跟踪

提出了一种基于局部区域内的机器人和目标密度的多机器人分布式控制算法。运用这种算法,多机器人在跟踪多目标过程中,每个机器人能在每次调整速度和方向之前,根据最新的通讯信息和感知信息,计算通讯范围内相邻机器人和感知范围内移动目标的数量及位置坐标,然后按照所受人工势场(APF)力的合力决定下一次移动的新位置坐标,这个合力乘以一个基于机器人和目标密度的可变加权系数,以适应不同状态,然后按照动力学和运动学方程计算速度及方向。通过仿真实验对该算法进行了验证,实验结果表明,多机器人可以根据多目标的移动进行自主协作跟踪观测,算法的性能优于无可变加权系数的APF法和静态感知节点(SN)法。此方法对于实时监测和目标跟踪具有实际应用意义。     

红外弱小目标检测跟踪并行处理系统

针对红外弱小目标的实时检测与跟踪,本文设计了一套基于4×ADSP TS201S-600处理器的多DSP并行处理系统.该系统拥有LVDS数字视频输入输出接口,利用TS201S的链路口构建了松耦合并行处理系统,支持多DSP间的两两交叉互联和板级互联,定点运算的峰值速度可达19.2 GMAC@16 bit/s,浮点运算的峰值速度可达14.4 GFLOPS.实验结果表明,该并行处理系统具有高实时性、良好的适用性和扩展性等特点,可以实现对大画面,高帧频的红外弱小目标实时检测跟踪.     

Hierarchical active shape model with motion prediction for real-time tracking of non-rigid objects

Video tracking systems generally deal with non-rigid objects with various shapes and sizes. This often results in a poor match of an initial model with the actual input shape, and consequently causes the failure of tracking. The robustness of the active shape model (ASM) enables video tracking systems to deal with such unpredictable inputs. The iterative nature of the ASM, however, makes real-time implementation difficult. A novel ASM-based real-time tracking method with particular relevance to non-rigid objects is proposed. The proposed tracking system adopts a hierarchical approach to reduce computational loads and increase immunity to noise. In order to make the system operate in real-time, a novel prediction approach is proposed that significantly reduces the number of iterations. In the sequential images, the initial feature points have been estimated using a block-matching algorithm. Kalman filtering has also been applied for increasing accuracy of the motion estimation. The proposed hierarchical, prediction-based approach is proven to outperform the existing methods in the sense of both tracking performance and convergence speed     

Model Discrimination for Nonlinear Regression Models

Wireless sensor networks (WSNs) are becoming important tools in various tasks, including monitoring and tracking of spatially occurring phenomena. These networks offer the capability of densely covering a large area, but at the same time are constrained by the limiting sensing, processing and power capabilities of their sensors. To complete the task at hand, the information collected by the sensor nodes needs to be appropriately fused. In this article we study the problems of estimating the location of a target and estimating its signal intensity. The proposed algorithms are based on the local vote decision fusion (LVDF) mechanism, where sensors first correct their original decisions using decisions of neighboring sensors. These corrected decisions are more accurate and robust and improve detection; however, they are correlated, which makes maximum likelihood estimation intractable. We adopt a pseudolikelihood formulation and examine several variants of localization and signal estimation algorithms based on original and corrected decisions using direct optimization methods, as well as an EM approach. Uncertainty assessments about the parameters of interest are provided using a parametric bootstrap technique. An extensive simulation study of the developed algorithms, along with several benchmarks, establishes the overall superior performance of the LVDF-based algorithms, especially in low signal-to-noise ratio environments. Extensions to tracking moving targets and localizing multiple targets also are considered.     

形心和匹配跟踪算法的改进

从数字图象处理方法着手介绍了一些有效的跟踪算法 ,分析了这些算法的优缺点 ,并根据工程应用的要求 ,对这些算法进行了改进 ,以满足工程要求。实验表明这些改进后的算法对复杂背景中目标的跟踪性能有较大提高。     

一种用于电视跟踪系统中的多模式数据融合方法

基于单传感器多模式数据融合的思想,设计了一种联合卡尔曼滤波器。通过卡尔曼滤波分 散处理多个模式跟踪结果数据,再将处理结果进行全局融合得出跟踪最终结果。仿真和室内实验都表明,这种方法用于实时电视跟踪系统中可有效提高跟踪精度,融合后的均方误差是融合前的0.05倍,具有一定的容错能力,运算量小,且易于实时实现。     

Analytic performance prediction of feature-aided global nearest neighbour algorithm in dense target scenarios

An analytic performance prediction method for the feature-aided global nearest neighbour tracking algorithm in multi-target tracking (MTT) scenarios is proposed. The approach serves as an alternative to the costly Monte Carlo simulation method. In MTT, evaluation of interference among multiple targets remains a crucial issue on tracking performance study. This issue is investigated in dense target scenarios with feature information and unrestrictive motion. Analytic expressions are developed for tracking performance in terms of the probability of correct association and estimation accuracy. Feature information of targets is incorporated in the formulation which provides us an insight on how the tracking performance is impacted by features. In the derivations, a series of simplification assumptions are made and the results are not intended to be used directly in practical tracking applications. The major contribution of the paper is to provide a theoretical exploration and a methodology for analytic performance prediction of MTT.     

Cluster-based centralized data fusion for tracking maneuvering targets using interacting multiple model algorithm

The interacting multiple model (IMM) algorithm has proved to be useful in tracking maneuvering targets. Tracking accuracy can be further improved using data fusion. Tracking of multiple targets using multiple sensors and fusing them at a central site using centralized architecture involves communication of large volumes of measurements to a common site. This results in heavy processing requirement at the central site. Moreover, track updates have to be obtained in the fusion centre before the next measurement arrives. For solving this computational complexity, a cluster-based parallel processing solution is presented in this paper. In this scheme, measurements are sent to the data fusion centre where the measurements are partitioned and given to the slave processors in the cluster. The slave processors use the IMM algorithm to get accurate updates of the tracks. The master processor collects the updated tracks and performs data fusion using ‘weight decision approach’. The improvement in the computation time using clusters in the data fusion centre is presented in this paper     

基于卷积网络的目标跟踪应用研究

本文针对目标跟踪应用,提出了基于Siamese-FC跟踪网络的改进卷积网络Siamese-MF,意在更进一步提升跟踪速度和准确性,满足目标跟踪的工程应用需求。对于跟踪网络,考虑速度和精度的权衡,减少计算量,增加卷积特征的感受野是改进跟踪网络的速度和精度的方向。在卷积网络结构上面进行改进结构创新,改进主要集中为两点:1)引入特征融合,丰富特征;2)引入空洞卷积,减少计算量的同时增强感受野。Siamese-MF算法实现了对于复杂场景目标的实时准确跟踪,在公开数据集OTB上测试速度达到平均76 f/s,跟踪成功率的均值达到0.44,而跟踪稳定性的均值达到0.61,实时性、准确性和稳定性均提升,满足目标实时跟踪应用。     

Real-Time Path Planning Based on the Situation Space of UCAVs in a Dynamic Environment

This paper aims to find a reliable, collision-free path in a dynamic environment for highly maneuverable unmanned combat air vehicles (UCAVs). Given the real-time nature of the operational scenario, quick and adaptable reactions of UCAVs are necessary for updates in situational awareness. Therefore, we propose a three dimensional (3D) path planning approach based on the situational space to provide the tactical requirements of UCAVs for tracking targets and avoiding collisions. First, to ensure reliable nonlinear measurements, the interacting multiple model (IMM) algorithm based on a cubature Kalman filter (CKF) is chosen for the tracking and prediction algorithm. A constraint reference frame combining the kinematic model of constant acceleration (CA) is developed to solve the problem of arrival point generation. Second, by analyzing the relative motion between the UCAV and the moving objects, we define the situation space and give the corresponding calculation method. In tracking the moving target, the guidance vector contains the fusion information of displacement and velocity. At the same time, taking advantage of the one-step situation space as the judgment of the threat, we further plan the collision avoidance strategy. Third, as the safety in a practically reachable trajectory of the UCAV possesses the absolute priority, the collision avoidance acceleration accounts for this dominant factor in path planning. Simulations and experimental results prove that the proposed approach can plan a smooth and flyable path in 0.008 s under the premise of soft-landing target tracking.