A Distributed Particle Filter Applied in Single Object Tracking

Visual object-tracking is a fundamental task applied in many applications of computer vision. Particle filter is one of the techniques which has been widely used in object tracking. Due to the virtue of extendability and flexibility on both linear and non-linear environments, various particle filter-based trackers have been proposed in the literature. However, the conventional approach cannot handle very large videos efficiently in the current data intensive information age. In this work, a parallelized particle filter is provided in a distributed framework provided by the Hadoop/Map-Reduce infrastructure to tackle object-tracking tasks. The experiments indicate that the proposed algorithm has a better convergence and accuracy as compared to the traditional particle filter. The computational power and the scalability of the proposed particle filter in single object tracking have been enhanced as well.

3D Single Object Tracking with Multi-View Unsupervised Center Uncertainty Learning

Center point localization is a major factor affecting the performance of 3D single object tracking.Point clouds themselves are a set of discrete points on the local surface of an object,and there is also a lot of noise in the labeling.Therefore,directly regressing the center coordinates is not very reasonable.Existing methods usually use volumetric-based,point-based,and view-based methods,with a relatively single modality.In addition,the sampling strategies commonly used usually result in the loss of object information,and holistic and detailed information is beneficial for object localization.To address these challenges,we propose a novel Multi-view unsupervised center Uncertainty 3D single object Tracker(MUT).MUT models the potential uncertainty of center coordinates localization using an unsupervised manner,allowing the model to learn the true distribution.By projecting point clouds,MUT can obtain multi-view depth map features,realize efficient knowledge transfer from 2D to 3D,and provide another modality information for the tracker.We also propose a former attraction probability sampling strategy that preserves object information.By using both holistic and detailed descriptors of point clouds,the tracker can have a more comprehensive understanding of the tracking environment.Experimental results show that the proposed MUT network outperforms the baseline models on the KITTI dataset by 0.8%and 0.6%in precision and success rate,respectively,and on the NuScenes dataset by 1.4%,and 6.1%in precision and success rate,respectively.The code is made available at https://github.com/abchears/MUT.git.

A Sensor Network Coverage Planning Based on Adjusted Single Candidate Optimizer

Wireless sensor networks(WSNs)are widely used for various practical applications due to their simplicity and versatility.The quality of service in WSNs is greatly influenced by the coverage,which directly affects the monitoring capacity of the target region.However,low WSN coverage and uneven distribution of nodes in random deployments pose significant challenges.This study proposes an optimal node planning strategy for net-work coverage based on an adjusted single candidate optimizer(ASCO)to address these issues.The single candidate optimizer(SCO)is a metaheuristic algorithm with stable implementation procedures.However,it has limitations in avoiding local optimum traps in complex node coverage optimization scenarios.The ASCO overcomes these limitations by incorporating reverse learning and multi-direction strategies,resulting in updated equations.The performance of the ASCO algorithm is compared with other algorithms in the literature for optimal WSN node coverage.The results demonstrate that the ASCO algorithm offers efficient performance,rapid convergence,and expanded coverage capabilities.Notably,the ASCO achieves an archival coverage rate of 88%,while other approaches achieve coverage rates below or equal to 85%under the same conditions.

Optical encryption of multiple three-dimensional objects based on multiple interferences and single-pixel digital holography

We present an optical encryption method of multiple three-dimensional objects based on multiple interferences and single-pixel digital holography. By modifying the Mach-Zehnder interferometer, the interference of the multiple objects beams and the one reference beam is used to simultaneously encrypt multiple objects into a ciphertext. During decryption, each three-dimensional object can be decrypted independently without having to decrypt other objects. Since the single- pixel digital holography based on compressive sensing theory is introduced, the encrypted data of this method is effectively reduced. In addition, recording fewer encrypted data can greatly reduce the bandwidth of network transmission. Moreover, the compressive sensing essentially serves as a secret key that makes an intruder attack invalid, which means that the system is more secure than the conventional encryption method. Simulation results demonstrate the feasibility of the proposed method and show that the system has good security performance.

Analysis of Object Depth Effects on Accuracy of Dimensional Shape in X and Y Directions Using Single Non-metric Image

In general, to reconstruct the accurate shape of buildings, we need at least one stereomodel （two photographs） for each building. In most cases, however, only a single non-metric photograph is available, which is usually obtained either by an amateur, such as a tourist, or from a newspaper or a post card. To evaluate the validity of 3D reconstruction from a single non-metric image, this study analyzes the effects of object depth on the accuracy of dimensional shape in X and Y directions using a single non-metric image by means of simulation technique, as this was considered to be, in most cases, a main source of data acquisition in recording and documenting buildings.

A Genetic Algorithm for Single Machine Scheduling with Fuzzy Processing Time and Multiple Objectives

In this paper, by considering the fuzzy nature of the data in real-life problems, single machine scheduling problems with fuzzy processing time and multiple objectives are formulated and an efficient genetic algorithm which is suitable for solving these problems is proposed. As illustrative numerical examples, twenty jobs processing on a machine is considered. The feasibility and effectiveness of the proposed method have been demonstrated in the simulation.

Searching for an Optimized Single-objective Function Matching Multiple Objectives with Automatic Calibration of Hydrological Models

In the calibration of hydrological models, evaluation criteria are explicitly and quantitatively defined as single-or multi-objective functions when utilizing automatic calibration approaches.In most previous studies, there is a general opinion that no single-objective function can represent all important characteristics of even one specific hydrological variable(e.g., streamflow).Thus hydrologists must turn to multi-objective calibration.In this study, we demonstrated that an optimized single-objective function can compromise multi-response modes(i.e., multi-objective functions) of the hydrograph, which is defined as summation of a power function of the absolute error between observed and simulated streamflow with the exponent of power function optimized for specific watersheds.The new objective function was applied to 196 model parameter estimation experiment(MOPEX) watersheds across the eastern United States using the semi-distributed Xinanjiang hydrological model.The optimized exponent value for each watershed was obtained by targeting four popular objective functions focusing on peak flows, low flows, water balance, and flashiness, respectively.Results showed that the optimized single-objective function can achieve a better hydrograph simulation compared to the traditional single-objective function Nash-Sutcliffe efficiency coefficient for most watersheds, and balance high flow part and low flow part of the hydrograph without substantial differences compared to multi-objective calibration.The proposed optimal single-objective function can be practically adopted in the hydrological modeling if the optimal exponent value could be determined a priori according to hydrological/climatic/landscape characteristics in a specific watershed.

Modeling and Analysis of Single Machine Scheduling Based on Noncooperative Game Theory

Considering the independent optimization requirement for each demander of modernmanufacture, we explore the application of noncooperative game in production scheduling research,and model scheduling problem as competition of machine resources among a group of selfish jobs.Each job has its own performance objective. For the single machine, multi-jobs and non-preemptivescheduling problem, a noncooperative game model is established. Based on the model, many prob-lems about Nash equilibrium solution, such as the existence, quantity, properties of solution space,performance of solution and algorithm are discussed. The results are tested by numerical example.

Single-Channel Speech Enhancement Using Critical-Band Rate Scale Based Improved Multi-Band Spectral Subtraction

This paper addresses the problem of single-channel speech enhancement in the adverse environment. The critical-band rate scale based on improved multi-band spectral subtraction is investigated in this study for enhancement of single-channel speech. In this work, the whole speech spectrum is divided into different non-uniformly spaced frequency bands in accordance with the critical-band rate scale of the psycho-acoustic model and the spectral over-subtraction is carried-out separately in each band. In addition, for the estimation of the noise from each band, the adaptive noise estimation approach is used and does not require explicit speech silence detection. The noise is estimated and updated by adaptively smoothing the noisy signal power in each band. The smoothing parameter is controlled by a-posteriori signal-to-noise ratio (SNR). For the performance analysis of the proposed algorithm, the objective measures, such as, SNR, segmental SNR, and perceptual evaluations of the speech quality are conducted for the variety of noises at different levels of SNRs. The speech spectrogram and objective evaluations of the proposed algorithm are compared with other standard speech enhancement algorithms and proved that the musical structure of the remnant noise and background noise is better suppressed by the proposed algorithm.

结合Cokriging模型和单目标函数的随机模型修正

将Cokriging代理模型技术和单目标函数进行结合,提出了一种随机模型修正方法。该方法将不确定性模型修正问题转化为较简单的待修正参数统计特征的修正问题,能够在保证模型修正精度的同时,有效缓解模型修正中由于分步修正和多目标修正造成的计算成本高的问题。首先,假设待修正参数和响应均服从高斯分布,采用拉丁超立方抽样获取训练集样本,构造满足精度要求的Cokriging模型替代复杂的有限元模型参与迭代计算。然后,建立有限元模型计算响应统计特征与试验响应统计特征的加权残差目标函数,引入土狼优化算法最小化该单目标函数来得到待修正参数的统计特征。最后,通过二维和三维桁架结构验证了所提方法的可行性。

多/单目标优化转换下的作战任务重分配

协同作战任务分配技术是近年来军事领域的研究热点之一,以往的研究一般将任务分配划分为预分配和动态分配2个阶段并独立进行研究,但是这种做法忽略了2个阶段之间的内在联系。针对这一问题,以多目标优化下的任务预分配和单目标优化下的任务动态分配为切入点,提出一种多/单目标优化转换思想。在动态任务分配阶段,通过决策者在预分配阶段的选择获取其主观偏好,基于获取的主观偏好将多目标优化转化为单目标优化后,使用合同网协议完成单目标任务重分配。仿真结果证明了所提出的多/单目标优化转换思想的正确性及其在动态任务分配问题中的适用性。

基于多通道交叉注意力融合的三维目标检测算法

针对现有单阶段三维目标检测算法对点云下采样特征利用方式单一、特征对长程上下文信息的聚合程度无法满足算法性能提升需求的问题,本文提出了基于多通道交叉注意力融合的单阶段三维目标检测算法。首先,设计通道交叉注意力模块用于融合下采样特征,可基于交叉注意力机制在通道层面上增强多尺度特征对不同感受野下长程空间信息的表达能力;然后,提出级联特征激励模块,结合原始下采样特征对通道交叉注意力加权特征进行级联激励,提升算法对关键空间特征的学习能力。在公共自动驾驶数据集KITTI上进行了大量实验并与主流算法对比,本文算法作为单阶段目标检测算法,在车辆类别3个难度级别上的检测准确率分别为91.34%、79.85%和75.98%,较基线算法分别提升了4.83%、3.26%和3.32%。实验结果证明了本文算法及所提模块在三维目标检测任务上的有效性和先进性。

二层多目标随机规划逼近弱有效解集的上半收敛性

为研究二层多目标随机规划逼近问题的弱有效解与精确的弱有效解之间的逼近收敛性,针对上、下层都带有约束条件的一类多目标二层随机规划的逼近问题,构建了二层多目标随机规划逼近问题的弱有效解集的上半收敛性理论框架。即在假设下层反馈到上层的最优解集函数为凸函数的前提下,借助严格凸函数的性质,利用多目标随机规划的弱有效解可以表示成相应的单目标随机规划最优解集交集的结构特征,建立了二层多目标随机规划逼近弱有效解集的上半收敛性,提供了逼近方法求解二层多目标随机规划弱有效解集可以近似替代精确的弱有效解集的理论依据。

普适型GNSS双频数据的周跳探测与修复方法

通过分析宽巷相位减窄巷伪距和无几何距离组合周跳探测方法存在的问题,提出一种考虑GNSS卫星高度角因子且适用于普适型GNSS观测数据的周跳探测方法。在探测出周跳发生的历元位置后,利用基于空间搜索和目标函数最小准则的方法进行周跳修复,并选取安置在甘肃临夏公路旁一滑坡隐患处的普适型GNSS接收机所观测的不同高度角卫星数据对提出的周跳处理方法进行验证。结果表明,该周跳探测与修复方法能实现不同高度角卫星载波相位数据中周跳的准确探测与修复,并减弱GNSS数据观测误差对周跳检验量的影响,对普适型GNSS观测数据的周跳探测与修复具有可行性。

基于YOLOv5s的水稻种子逐粒检测装置研制

针对农机化智能检测场景下的水稻种子批量单粒活力检测的需求,结合计算机视觉目标检测技术,设计并研制气吸式种子逐粒检测装置。为解决水稻种子这一小目标的检测效果不佳和漏检问题,提出一种基于YOLOv5s的改进算法。在检测网络中设计上下文特征融合机制,利用新的特征融合方式增强网络多尺度空间和通道信息表征,并添加CA注意力模块,在此基础上改善网络特征融合性能。设计由排列装置、传送装置、检测装置、分选装置等组成的水稻种子逐粒分选线。选取中嘉早17、甬优538和中浙优10三种不同外形的水稻为试验对象,以吸嘴直径、转速、压力作为自变量,以吸空率、单粒吸附率和识别正确率为考核指标,对影响水稻种子检测效果进行试验研究和分析,对于三种不同形状的种子分别选取恰当的吸嘴直径、转速、压力能够有效地吸附并逐粒排列。试验结果表明:改进算法结合研制的系统能满足不同场景检测要求。

平面并联机器人离线PID控制优化研究

为了使平面并联机器人在闭环控制系统中具有更稳定的控制性能。对此,这里构建了平面五杆并联机器人动力学模型及其控制系统模型,给出了控制系统的增益矢量p。对机器人系统模型进行离线PID控制优化,通过控制增益来设计控制系统中的增益矢量p,从而实现非线性单目标动态优化(NLMODOP)。在NLMODOP中加入动态约束,采用带约束处理机制的差分进化(DE)算法求解平面并联机器人的非线性规划问题,进而处理不稳定的动态优化。对机器人模型中的五个连杆进行仿真实验,并对有无DE算法控制的仿真结果进行了比较。结果表明:相比于无DE算法,采用DE算法下的机器人系统模型的连杆跟踪位移基本无跟踪误差。说明基于差分进化算法的平面并联机器人离线PID控制优化具有较好的控制精度和跟踪性能。

基于多重注意力机制与响应融合的孪生单目标跟踪算法

针对孪生全卷积网络的单目标跟踪算法因无法提取到目标的高层语义特征和无法一次性集中关注并学习到目标的通道、空间及坐标特征导致在复杂场景下面临目标形变、姿态变化及背景干扰等挑战时,出现跟踪性能下降以及跟踪失败的问题,提出了一种基于多重注意力机制与响应融合的孪生网络单目标跟踪算法用来解决这一问题。在该算法中设计了小卷积核与跳层连接特征融合的深层骨干特征提取网络、改进型注意力机制及卷积互相关后的响应融合运算这3个模块用来提升该算法的跟踪性能,并通过消融实验验证了这3个模块的有效性。最后,经在OTB100基准数据集上测试,跟踪精确度达到了0.825,跟踪成功率达到了0.618。同时与其他先进算法进行对比,结果表明该算法不仅可以有效应对复杂场景下目标跟踪算法性能下降的问题,还可以在保证跟踪速度的前提下,进一步提高跟踪的精度。

基于YOLOv5m和CBAM-CPN的单分蘖水稻表型参数提取

为快速获取单分蘖水稻植株的形态结构和表型参数,该研究提出了一种基于目标检测和关键点检测模型相结合的骨架提取和表型参数获取方法。该方法基于目标检测模型生成穗、茎秆、叶片的边界框和类别,将所得数据分别输入到关键点检测模型检测各部位关键点,按照语义信息依次连接关键点形成植株骨架,依据关键点坐标计算穗长度、茎秆长度、叶片长度、叶片-茎秆夹角4种表型参数。首先,构建单分蘖水稻的关键点检测和目标检测数据集;其次,训练Faster R-CNN、YOLOv3、YOLOv5s、YOLOv5m目标检测模型,经过对比,YOLOv5m的检测效果最好,平均精度均值(mean average precision,mAP)达到91.17%;然后,应用人体姿态估计的级联金字塔网络(cascaded pyramid network,CPN)提取植株骨架,并引入注意力机制CBAM(convolutional block attention module)进行改进,与沙漏网络(hourglass networks,HN)、堆叠沙漏网络模型(stacked hourglass networks,SHN)和CPN模型相比,CBAM-CPN模型的预测准确率分别提高了9.68、8.83和1.06个百分点,达到94.75%,4种表型参数的均方根误差分别为1.06 cm、0.81 cm、1.25 cm和2.94°。最后,结合YOLOv5m和CBAM-CPN进行预测,4种表型参数的均方根误差分别为1.48、1.05、1.74cm和2.39°,与SHN模型相比,误差分别减小1.65 cm、3.43 cm、2.65 cm和4.75°,生成的骨架基本能够拟合单分蘖水稻植株的形态结构。所提方法可以提高单分蘖水稻植株的关键点检测准确率,更准确地获取植株骨架和表型参数,有助于加快水稻的育种和改良。

一种基于离散时域模型的单相PWM整流器控制参数多目标优化设计方法

单相PWM整流器一般采用电压外环比例积分、电流内环比例谐振的双闭环控制方法,在实现交流侧高功率因数的同时稳定直流侧输出电压,因此受到国内外学者的广泛研究。然而,在内外环控制参数设计阶段,传统设计方法一般根据经验选取控制参数,参数选取范围较大,且无法量化评估控制参数的时域控制效果,导致参数优化整定困难。为此,该文提出一种基于离散时域模型的内外环控制参数多目标优化设计方法。首先,采用传统设计方法确定电压外环比例、积分控制参数以及电流内环比例、谐振控制参数的安全设计空间;然后,基于离散时域模型将安全设计空间正映射至控制参数的时域性能空间,实现电流内环控制参数对交流侧功率因数以及电压外环控制参数对直流侧电压超调量的量化评估。在此基础上,通过进一步约束性能目标得到性能优化设计空间,考虑内外环控制参数的耦合效应,对内外环性能优化设计空间进行重新组合,再将其映射至性能空间,得到内外环控制目标帕累托最优前沿下的最优设计空间,从中选取内外环控制参数的优化设计点,在典型工况下实验测得的交流侧功率因数为0.981 26,直流输出电压超调量绝对值为1.37%,证明了所提优化设计方法的正确性。

不同目标函数对径流模拟结果的影响分析

水文模型是洪水模拟的重要工具,大部分模型包含一定数量的参数,因此,参数率定是提高水文模型径流模拟精度、保证模型结果准确性关键之一,而目标函数作为参数率定过程的寻优标准,目前缺乏详尽分类体系为目标函数的选取提供明确选择依据。为探明不同类型目标函数对水文模型参数率定结果和径流模拟精度的影响,综合考虑纳什效率系数、Kling-Gupta系数、洪量相对误差等多种不同类型目标,构建一种包含高流量目标、低流量目标、洪量目标、洪水过程目标及加权目标的目标函数分类体系,结合当前水文预报情报规范的要求,为目标函数选择提供一定标准;进一步,建立新安江水文模型,引入SCE-UA、NSGA-Ⅱ优化算法,为不同算法设置多种目标函数选择方案,以汉江流域牧马河为例进行径流模拟,分析不同目标函数方案和优化算法对径流模拟精度的影响。结果表明:采用单一目标函数,可提高该目标函数所体现水文要素的模拟精度;选取合适的目标函数是提升径流模拟精度的关键,在现有规范的要求下,选取洪水过程目标和高流量目标是保证水文模型精度的最佳选择;Kling-Gupta系数与纳什效率系数相比,其在洪量模拟精度上表现更佳,是洪水过程目标的更优选择;加权单目标与多目标方法的模拟结果在高流量部分存在差异。