基于TLD算法的加油锥套跟踪

针对自主空中加油对接阶段锥套跟踪问题,提出了一种基于tracking learning detection (TLD)的锥套跟踪算法。该算法将加油锥套的跟踪任务分解成跟踪、学习、检测3个部分。跟踪模块在LK光流法的基础上添加跟踪失败自检测,筛选出好的跟踪点,跟踪加油锥套;检测模块构建级联分类器,对滑动窗遍历得到的图像块进行分类并返回含有目标的图像块,融合跟踪模块的跟踪框,给出最终跟踪结果;学习模块引入P N约束修正错误样本并学习更新检测模块。利用Creator/Vega Prime软件对空中加油进行视景仿真,在视景仿真视频上测试锥套跟踪算法。结果表明:TLD算法跟踪加油锥套成功率达95.5%,处理每帧平均耗时31.4 ms,能够满足加油锥套跟踪鲁棒性、准确率、实时性的要求。     

MS-KF融合算法用于锥套跟踪

针对无人机自主空中加油过程中锥套跟踪,提出一种均值漂移-卡尔曼滤波(mean shift-Kalman filter, MS-KF) 融合算法。分析了基于均值漂移算法的锥套目标模型、相似性度量、锥套目标定位的锥套定位原理;引入卡尔曼滤波器对锥套运动状态进行预测,将锥套运动信息融合到均值漂移算法中,以保证锥套跟踪算法的稳定性和鲁棒性;给出了MS-KF融合算法用于锥套识别跟踪的流程;搭建了锥套跟踪半物理实验验证系统,分别进行MS-KF融合算法用于锥套跟踪的半物理实验验证及数值仿真分析。实验结果表明:MS-KF融合算法可以对锥套精确定位跟踪,无人机3个轴向的跟踪误差保持在0.3 m的范围内,保证了无人机自主空中加油的顺利进行。     

基于改进UKF的加油机位姿预测方法

针对无人机自主空中加油保持阶段加油机位姿跟踪精度不高的问题,提出了一种改进UKF（无损卡尔曼滤波）预测方法。建立了视觉导航系统模型,利用Harris算法检测角点,并用RANSAC（随机序列一致性）算法进行角点匹配。将历史预测数据引入当前时刻UKF预测值,并通过匹配角点所得姿态观测值对改进UKF预测值进行修正,从而实现加油机姿态的高精度预测。仿真结果表明,改进UKF在遭遇突发强干扰时姿态预测性能明显优于标准UKF,所预测误差小于5.8%,满足空中加油精度要求。该算法避免了强干扰引发的预测出错,有效抑制了突发干扰。     

高空无人机自动加油对接过程轨迹控制研究

针对无人机在高空自动加油对接过程中容易出现偏差的问题,提出了无人机空中自动加油精准对接的优化控制方案,研究并分析了影响加油锥管与受油接口快速准确对接的主要因素,通过运动轨迹生成器对浮锚加油锥套的运动轨迹进行模拟,无人机的自动飞行控制系统根据轨迹跟踪曲线对无人机的飞行姿态进行调整,引导无人机的受油接口与加油锥管进行精确对接,整个飞行轨迹跟踪控制过程优化了复杂高空环境中无人机加油对接的精准性。通过仿真实验表明,本高空无人机加油对接引导方法提高了空中管口对接的速度与抗干扰能力,可以满足各种复杂高空环境的无人机加油任务。     

硬管式无人机AAR双目视觉导航算法研究

针对硬管式无人机自主空中加油近距编队阶段的相对位置和姿态估计问题,研究了基于双目视觉的相对位姿估计算法。该算法采用Harris方法提取特征点,并对其进行快速匹配,通过Sampson方法三维重构获得特征点在摄像机坐标系下的三维坐标,以重构误差平方和最小为准则建立目标函数,利用单位四元数法求解位姿参数。最后利用仿真平台验证双目视觉位姿估计算法的有效性。结果表明:相对位置误差低于0.1 m,相对姿态误差小于0.5°,其精度满足自主空中加油相对导航性能要求。     

基于影像实时处理的加油对接段辅助对准技术北大核心CSCD

针对空中加油试飞中加油对接困难的问题,设计了一种基于影像实时处理的加油对接段辅助对准系统,通过影像测量技术计算受油头与加油锥套中心的精确相对位置,实现位置参数与视频画面实时同步显示,用于空中加受油对接辅助对准。并对复杂光学条件下锥套跟踪技术、基于约束的像机标定技术和加受油组件相对位置实时测量等关键技术进行研究。实验结果表明,该算法可实现空中复杂环境下锥套图像快速、稳定识别与跟踪;采用双目视觉前方交会测量实时计算加油锥套与受油头相对位置,与事后处理结果对比分析,精度优于0.1 m,可辅助飞行员进行空中加油对接操作,提高加受油对接成功率。     

Pose estimation for UAV aerial refueling with serious turbulences based on extended Kalman filter

In recent years, many pose estimation algorithms were developed, and have been successfully applied to solve unmanned aerial vehicle (UAV) aerial refueling pose estimation problems. This paper mainly focuses on solving this problem under serious turbulences circumstance. The extended Kalman filter is a set of mathematical equations to estimate the state of a process, which is able to support estimations of past, present, and even future states. In reference to previous papers and some simulations, we build up the noise models of refueling boom and atmospheric turbulence. Then, an extend Kalman filter is adopted to solve the pose estimation problem in UAV aerial refueling with serious turbulences. The experimental results demonstrate the feasibility and effectiveness of our proposed approach.     

基于改进SSD的车辆小目标检测方法

地面车辆目标检测问题中由于目标尺寸较小,目标外观信息较少,且易受背景干扰等的原因,较难精确检测到目标。围绕地面小尺寸目标精准检测的问题,从目标特征提取的角度提出了一种特征融合的子网络。该子网络引入了重要的局部细节信息,有效地提升了小目标检测效果。针对尺度、角度等的变换问题,设计了基于融合层的扩展层预测子网络,在扩展层的多个尺度空间内匹配目标,生成目标预测框对目标定位。在车辆小目标VEDAI(vehicle detection in aerial imagery)数据集上的实验表明,算法保留传统SSD(single-shot multibox detector)检测速度优势的同时,在精度方面有了明显提升,大幅提升了算法的实用性。     

基于图像的空中加油软管平衡拖曳位置测量方法

针对飞行条件下加油软管平衡拖曳位置的获取难题,设计了一套基于图像的空中加油软管平衡拖曳位置测量方法。通过在加油机上加装多路高清影像测量系统获取空中加油过程软管的运动图像,并对多像机同步采集、大景深像机标定、无参考点加油锥套识别与跟踪等关键技术进行研究,实现了加油软管平衡拖曳位置的精确测量。仿真实验结果表明,采用该方案测量精度优于5 cm,满足飞行试验精度要求。     

像素重要性测量特征下的侧扫声呐目标检测

研究了在较低信噪比下,在保证检测概率的前提下尽量降低虚警概率的目标检测,提出了一种针对特定目标的两阶段筛选算法.第一阶段中,首先使用阈值分割出有效点,并定义了一种新的像素重要性测量特征用于初步筛选目标。即通过有效像素点之间的距离来赋以高斯分布的权值,当前像素重要性的值定义为剩余有效点的距离加权和,具有较高的像素重要性值的聚集性强的区域内像素点会被定位出来。第二阶段,使用卷积神经网络分类器排除虚假目标.在实验中,使用近期无人潜器获得的海底数据,召回率与虚警概率分别达到90.39%与2.39%,证明了其相比声呐目标检测主流算法有更好的检测能力。      

Flocking in Multi-Agent Systems with Multiple Virtual Leaders Based Only on Position Measurements

Most existing flocking algorithms assume one single virtual leader and rely on information on both relative positions and relative velocities among neighboring agents. In this paper, the problem of controlling a flock of mobile autonomous agents to follow multiple virtual leaders is investigated by using only position information in the sense that agents with the same virtual leader asymptotically attain the same velocity and track the corresponding virtual leader based on only position measurements. A flocking algorithm is proposed under which every agent asymptotically attains its desired velocity, collision between agents can be avoided, and the final tight formation minimizes all agents' global potentials. A simulation example is presented to verify and illustrate the theoretical results.     

结合机载LiDAR数据与航空可见光影像多层次规则分类建筑物变化检测

建议了一种结合Lidar点云与航空可见光影像的建筑物变化检测新方法,利用多层次规则分类算法解决这两种异元异构数据间建筑物变化检测难题。并建议了一种结合面积阈值的形态学后处理方法,从而形成一套完整的处理流程,可应用于实际生产。最终,利用中国吉林省长春市2010年机载LiDAR点云数据和2009年高分辨率航空影像对该方法的有效性进行了评价,通过与基于支持向量机(SVM)面向对象分类的建筑物变化检测算法比较,进一步对本研究建议的方法进行了验证与分析。结果显示,此方法效果理想,其精度优于基于SVM面向对象分类的建筑物变化检测方法。Kappa系数达到0.90,correctness达到0.87。     

无人飞行器视觉SLAM特征检测和匹配算法研究

近几年图像局部特征检测和描述在机器人视觉中得到了广泛的应用,鲁棒的、快速且高精度的视觉特征检测和描述算法对飞行器进行实时的位姿估计和地图构建具有决定性意义。本文针对四旋翼无人飞行器平台的RGB-D传感器同时定位与地图构建(SLAM),讨论FAST、STAR、SIFT和SURF等检测算法和ORB、FREAK和SURF等匹配描述符的性能,对不同的特征算法进行对比评估出最合适的特征检测方法和匹配描述符。最后,基于Eclipse与OpenCV平台进行了实验,实验结果表明FAST检测和FREAK描述符比其他方法更适用于四旋翼飞行器在板视觉SLAM,且能基本满足实时性。     

Convex relaxation based sparse algorithm for hyperspectral target detection

Target detection in hyperspectral images is an important task. In this paper, we propose a sparsity based algorithm for target detection in hyperspectral images. In sparsity model, each hyperspectral pixel is represented by a linear combination of a few samples from an overcomplete dictionary, and the weighted vector for such reconstruction is sparse. This model has been applied in hyperspectral target detection and solved with several greedy algorithms. As conventional greedy algorithms may be trapped into a local optimum, we consider an alternative way to regularize the model and find a more accurate solution to the model. The proposed method is based on convex relaxation technique. The original sparse representation problem is regularized with a properly designed weighted ?₁ minimization and effectively solved with existing solver. The experiments on synthetic and real hyperspectral data suggest that the proposed algorithm outperforms the classical sparsity-based detection algorithms, such as Simultaneous Orthogonal Matching Pursuit (SOMP) and Simultaneous Subspace Pursuit (SSP) and conventional ?₁ minimization.     

Joint terminal-AP association and power allocation for NOMA-enabled space–air–ground integrated networks

This paper considers a space–air–ground integrated network (SAGIN) to provide network access services for aerial and terrestrial terminals. The non-orthogonal multiple access (NOMA) is used for improving spectral efficiency in the uplink transmission between terminals and access points (APs) in SAGIN. A sum rate maximization optimization problem is formulated by optimizing terminal-AP association and power allocation, while simultaneously satisfying the constraints of transmit power, network coverage characteristics, and quality-of-service (QoS) requirements of both aerial and terrestrial terminals. To deal with the formulated mixed integer nonlinear programming (MINLP) optimization problem, we first decouple it into separated terminal-AP association and power allocation problems. Then, we adopt the Q-learning algorithm to solve the terminal-AP association subproblem. Based on the obtained terminal-AP association solution, an iterative power allocation algorithm is developed by exploiting the Lagrange dual method. Moreover, the computational complexity of the proposed algorithm is further analyzed. Simulation results demonstrate that, compared with other schemes, our proposed algorithm can achieves a better performance in terms of the achievable sum rate, average achievable rate, and outage probability.     

An Adaptive Deblurring Vehicle Detection Method for High-Speed Moving Drones: Resistance to Shake

Vehicle detection is an essential part of an intelligent traffic system, which is an important research field in drone application. Because unmanned aerial vehicles (UAVs) are rarely configured with stable camera platforms, aerial images are easily blurred. There is a challenge for detectors to accurately locate vehicles in blurred images in the target detection process. To improve the detection performance of blurred images, an end-to-end adaptive vehicle detection algorithm (DCNet) for drones is proposed in this article. First, the clarity evaluation module is used to determine adaptively whether the input image is a blurred image using improved information entropy. An improved GAN called Drone-GAN is proposed to enhance the vehicle features of blurred images. Extensive experiments were performed, the results of which show that the proposed method can detect both blurred and clear images well in poor environments (complex illumination and occlusion). The detector proposed achieves larger gains compared with SOTA detectors. The proposed method can enhance the vehicle feature details in blurred images effectively and improve the detection accuracy of blurred aerial images, which shows good performance with regard to resistance to shake.     

基于多检测器最大熵融合的多通道光谱图像异常检测

提出了一种多检测器最大熵融合的多通道光谱图像异常检测算法.选择多个不同的异常检测器,并利用自适应窗宽非参核密度估计方法估计其各自的输出分布,保留了多通道光谱图像数据的“长尾”特性,且避免了先验模型假设带来的模型误差.将各原始检测器的输出投影到具有标准正态边缘分布的变换空间中,利用变换空间中模型化的最大熵融合规则实现多检测器的决策级最优概率融合.在原数据空间通过似然函数的检验完成多通道光谱图像的目标检测.利用机载EPS-A航拍多通道光谱图像进行了实验,实验结果表明了算法的有效性.     

基于稳健背景子空间的高光谱图像异常检测

在RX算法中,局部背景协方差矩阵估计会由于背景受到异常像元的“污染”而不能准确反映背景分布,从而导致检测性能下降.针对这一问题,提出一种基于稳健背景子空间的异常检测算法.利用空间秩深度度量背景中每个样本相对于整个背景样本分布空间的位置,将“游离”于背景分布空间之外的样本看作是潜在的异常样本,并将其映射到背景分布空间之内.在此基础上,通过估计背景的协方差矩阵,利用主成分分析构造能更精确地刻画背景的子空间,在该子空间进行了基于马氏距离的检测异常.模拟和真实数据验证了该算法的有效性.     

基于灰色预测和KID离群点检测的AUV故障诊断

提高故障诊断能力对于确保水下机器人AUV系统的稳定运行具有重要意义。针对水下机器人推进器系统,提出一种基于离群点检测的AUV故障检测方法。首先,将传感器采集的数据进行灰色预测处理;然后,提出了一种结合K-mean和DBSCAN的改进迭代聚类(Iterative K-mean DBSCAN,IKD)算法进行离群点检测;最后,与K-mean和DBSCAN算法相比,仿真实验结果表明基于灰色预测和KID离群点检测算法的故障检测准确率高,能够有效地实现水下机器人AUV的无监督故障诊断。     

Joint detection and tracking of size-varying infrared targets based on block-wise sparse decomposition

The high variability of target size makes small target detection in Infrared Search and Track (IRST) a challenging task. A joint detection and tracking method based on block-wise sparse decomposition is proposed to address this problem. For detection, the infrared image is divided into overlapped blocks, and each block is weighted on the local image complexity and target existence probabilities. Target-background decomposition is solved by block-wise inexact augmented Lagrange multipliers. For tracking, label multi-Bernoulli (LMB) tracker tracks multiple targets taking the result of single-frame detection as input, and provides corresponding target existence probabilities for detection. Unlike fixed-size methods, the proposed method can accommodate size-varying targets, due to no special assumption for the size and shape of small targets. Because of exact decomposition, classical target measurements are extended and additional direction information is provided to improve tracking performance. The experimental results show that the proposed method can effectively suppress background clutters, detect and track size-varying targets in infrared images.