基于PointCloudTransformer和优化集成学习的三维点云分类

针对三维点云的不规则性和无序性所导致的难于提取特征并进行分类的问题,提出了一种融合深度学习和集成学习的三维点云分类方法。首先,训练深度学习点云分类网络PointCloudTransformer,并使用主干网络提取点云特征,进而训练基分类器,获得基分类器集合;然后,针对集成学习算法设计基分类器选择模型,模型的优化目标为基分类器组合的差异性和平均总体精度。为了降低集成规模,本文基于增强后的白鲸优化算法提出了二元多目标白鲸优化算法,并使用该算法优化基分类器选择模型,获得集成剪枝方案集合;最后,采用多数投票法集成每个基分类器组合在测试集点云特征上的分类结果,获得最优基分类器组合,从而构建基于多目标优化剪枝的集成学习点云分类模型。在点云分类数据集上的实验结果表明,本文方法使用了更小的集成规模,获得了更高的集成精度,能够对多类别三维点云进行准确分类。

Multi-Objective Optimization Algorithm for Grouping Decision Variables Based on Extreme Point Pareto Frontier

The existing algorithms for solving multi-objective optimization problems fall into three main categories:Decomposition-based,dominance-based,and indicator-based.Traditional multi-objective optimization problemsmainly focus on objectives,treating decision variables as a total variable to solve the problem without consideringthe critical role of decision variables in objective optimization.As seen,a variety of decision variable groupingalgorithms have been proposed.However,these algorithms are relatively broad for the changes of most decisionvariables in the evolution process and are time-consuming in the process of finding the Pareto frontier.To solvethese problems,a multi-objective optimization algorithm for grouping decision variables based on extreme pointPareto frontier(MOEA-DV/EPF)is proposed.This algorithm adopts a preprocessing rule to solve the Paretooptimal solution set of extreme points generated by simultaneous evolution in various target directions,obtainsthe basic Pareto front surface to determine the convergence effect,and analyzes the convergence and distributioneffects of decision variables.In the later stages of algorithm optimization,different mutation strategies are adoptedaccording to the nature of the decision variables to speed up the rate of evolution to obtain excellent individuals,thusenhancing the performance of the algorithm.Evaluation validation of the test functions shows that this algorithmcan solve the multi-objective optimization problem more efficiently.

An Efficient Reliability-Based Optimization Method Utilizing High-Dimensional Model Representation and Weight-Point Estimation Method

The objective of reliability-based design optimization(RBDO)is to minimize the optimization objective while satisfying the corresponding reliability requirements.However,the nested loop characteristic reduces the efficiency of RBDO algorithm,which hinders their application to high-dimensional engineering problems.To address these issues,this paper proposes an efficient decoupled RBDO method combining high dimensional model representation(HDMR)and the weight-point estimation method(WPEM).First,we decouple the RBDO model using HDMR and WPEM.Second,Lagrange interpolation is used to approximate a univariate function.Finally,based on the results of the first two steps,the original nested loop reliability optimization model is completely transformed into a deterministic design optimization model that can be solved by a series of mature constrained optimization methods without any additional calculations.Two numerical examples of a planar 10-bar structure and an aviation hydraulic piping system with 28 design variables are analyzed to illustrate the performance and practicability of the proposed method.

Optimization of the Use of Spherical Targets for Point Cloud Registration Using Monte Carlo Simulation

Registrations based on the manual placement of spherical targets are still being employed by many professionals in the industry.However,the placement of those targets usually relies solely on personal experience without scientific evidence supported by numerical analysis.This paper presents a comprehensive investigation,based on Monte Carlo simulation,into determining the optimal number and positions for efficient target placement in typical scenes consisting of a pair of facades.It demonstrates new check-up statistical rules and geometrical constraints that can effectively extract and analyze massive simulations of unregistered point clouds and their corresponding registrations.More than 6×10^(7) sets of the registrations were simulated,whereas more than IOO registrations with real data were used to verify the results of simulation.The results indicated that using five spherical targets is the best choice for the registration of a large typical registration site consisting of two vertical facades and a ground,when there is only a box set of spherical targets available.As a result,the users can avoid placing extra targets to achieve insignificant improvements in registration accuracy.The results also suggest that the higher registration accuracy can be obtained when the ratio between the facade-to-target distance and target-to-scanner distance is approximately 3:2.Therefore,the targets should be placed closer to the scanner rather than in the middle between the facades and the scanner,contradicting to the traditional thought. Besides,the results reveal that the accuracy can be increased by setting the largest projected triangular area of the targets to be large.

Optimal design of sintered Ce_9Nd_(21)Fe_(bal)B_1 magnets with a low-melting-point (Ce,Nd)-rich phase

A systemic investigation was done on the chemistry and crystal structure of boundary phases in sintered Ce9Nd21FebalB1 （wt%） magnets. Ce2Fe14B is believed to be more soluble in the rare-earth （RE）-rich liquid phase during the sintering process. Thus, the grain size and oxygen content were controlled via low-temperature sintering, resulting in high coercivity and maximum energy products. In addition, Ce formed massive agglomerations at the triple-point junctions, as confirmed by elemental mapping results. Transmission electron micros- copy （TEM） images indicated the presence of （Ce,Nd）Ox phases at grain boundaries. By controlling the composition and optimizing the preparation process, we successfully obtained Ce9Nd21FebalBx sintered magnets; the prepared magnets exhibited a residual induction, coerciv- ity, and energy product of 1.353 T, 759 kA/m, and 342 kJ/m3, respectively.

DETERMINATION OF PERMANENT OPTIMAL DATA POINTS AND AN EFFICIENT ALGORITHM FOR LAD PROBLEM

This paper gives a definition of permanent optimal data point of Least Absolute Deviation(LAD)problem.Some theoretical results on non-degenerate LAD problem are obtained.For computing LAD problem,an efficient,algorithm is given according to the idea of permanent optimal data point.Numerical experience shows that our algorithm is better than many of others,including the famous B R algorithm.

Optimal time point for the transplantation of neural stem cells induced to differentiate with retinoic acid

Previous studies have demonstrated that differentiated neural stem cells （NSCs） are more suitable for transplantation than non-differentiated NSCs. In this study, NSCs were expanded in vitro for two passages, induced with retinoic acid to differentiate, and harvested between 1 6 days later. They were subsequently cultured in artificial cerebrospinal fluid for an additional 3 days, dudng which their growth and morphology was monitored. NSCs induced for 4 days exhibited a peak rate of cells differentiating into neurons and robust growth. Our results indicate that the optimal time point for transplanting NSCs is following a 4-day period of induced differentiation.

Saddle Point Optimality Criteria of Interval Valued Non-Linear Programming Problem

The present paper aims to develop the Kuhn-Tucker and Fritz John criteria for saddle point optimality of interval-valued nonlinear programming problem.To achieve the study objective,we have proposed the definition of minimizer and maximizer of an interval-valued non-linear programming problem.Also,we have introduced the interval-valued Fritz-John and Kuhn Tucker saddle point problems.After that,we have established both the necessary and sufficient optimality conditions of an interval-valued non-linear minimization problem.Next,we have shown that both the saddle point conditions(Fritz-John and Kuhn-Tucker)are sufficient without any convexity requirements.Then with the convexity requirements,we have established that these saddle point optimality criteria are the necessary conditions for optimality of an interval-valued non-linear programming with real-valued constraints.Here,all the results are derived with the help of interval order relations.Finally,we illustrate all the results with the help of a numerical example.

Dynamics and Equilibria of N Point Charges on a 2D Ellipse or a 3D Ellipsoid

We consider the so-called Thomson problem which refers to finding the equilibrium distribution of a finite number of mutually repelling point charges on the surface of a sphere, but for the case where the sphere is replaced by a spheroid or ellipsoid. To get started, we first consider the problem in two dimensions, with point charges on circles (for which the equilibrium distribution is intuitively obvious) and ellipses. We then generalize the approach to the three-dimensional case of an ellipsoid. The method we use is to begin with a random distribution of charges on the surface and allow each point charge to move tangentially to the surface due to the sum of all Coulomb forces it feels from the other charges. Deriving the proper equations of motion requires using a projection operator to project the total force on each point charge onto the tangent plane of the surface. The position vectors then evolve and find their final equilibrium distribution naturally. For the case of ellipses and ellipsoids or spheroids, we find that multiple distinct equilibria are possible for certain numbers of charges, depending on the starting conditions. We characterize these based on their total potential energies. Some of the equilibria found turn out to represent local minima in the potential energy landscape, while others represent the global minimum. We devise a method based on comparing the moment-of-inertia tensors of the final configurations to distinguish them from one another.

Slope deformation partitioning and monitoring points optimization based on cluster analysis

The scientific and fair positioning of monitoring locations for surface displacement on slopes is a prerequisite for early warning and forecasting.However,there is no specific provision on how to effectively determine the number and location of monitoring points according to the actual deformation characteristics of the slope.There are still some defects in the layout of monitoring points.To this end,based on displacement data series and spatial location information of surface displacement monitoring points,by combining displacement series correlation and spatial distance influence factors,a spatial deformation correlation calculation model of slope based on clustering analysis was proposed to calculate the correlation between different monitoring points,based on which the deformation area of the slope was divided.The redundant monitoring points in each partition were eliminated based on the partition's outcome,and the overall optimal arrangement of slope monitoring points was then achieved.This method scientifically addresses the issues of slope deformation zoning and data gathering overlap.It not only eliminates human subjectivity from slope deformation zoning but also increases the efficiency and accuracy of slope monitoring.In order to verify the effectiveness of the method,a sand-mudstone interbedded CounterTilt excavation slope in the Chongqing city of China was used as the research object.Twenty-four monitoring points deployed on this slope were monitored for surface displacement for 13 months.The spatial location of the monitoring points was discussed.The results show that the proposed method of slope deformation zoning and the optimized placement of monitoring points are feasible.

Perceptual Optimization for Point-Based Point Cloud Rendering

Point-based rendering is a common method widely used in point cloud rendering.It realizes rendering by turning the points into the base geometry.The critical step in point-based rendering is to set an appropriate rendering radius for the base geometry,usually calculated using the average Euclidean distance of the N nearest neighboring points to the rendered point.This method effectively reduces the appearance of empty spaces between points in rendering.However,it also causes the problem that the rendering radius of outlier points far away from the central region of the point cloud sequence could be large,which impacts the perceptual quality.To solve the above problem,we propose an algorithm for point-based point cloud rendering through outlier detection to optimize the perceptual quality of rendering.The algorithm determines whether the detected points are outliers using a combination of local and global geometric features.For the detected outliers,the minimum radius is used for rendering.We examine the performance of the proposed method in terms of both objective quality and perceptual quality.The experimental results show that the peak signal-to-noise ratio(PSNR)of the point cloud sequences is improved under all geometric quantization,and the PSNR improvement ratio is more evident in dense point clouds.Specifically,the PSNR of the point cloud sequences is improved by 3.6%on average compared with the original algorithm.The proposed method significantly improves the perceptual quality of the rendered point clouds and the results of ablation studies prove the feasibility and effectiveness of the proposed method.

顾及邻域相关关系的改进PointNet++网络点云分类应用

机载扫描系统可以快速准确地获取大范围地物坐标,具有无可比拟的优势。机载点云具有其自身的特点,而采用PointNet++网络处理时未能依据机载点云的特点进行设计,其分类精度仍有待提高。鉴于此,提出了一种改进PointNet++网络,该网络将邻域点特征和邻域点间关系特征进行采集,并采用CRF算法对预测值进行全局优化用以增强空间一致性。通过选用ISPRS三维语义分割数据集对改进模型进行了测试,实验结果表明,改进PointNet++网络丰富了机载点云特征的描述,具有较高的分类精度。该成果可为三维地理信息提取提供有力保障。

采用混合策略联合优化的模糊C-均值聚类信息熵点云简化算法

针对传统聚类算法处理点云简化问题时精度低、耗时长且易丢失特征信息等问题,提出了一种基于动态精英自适应混合策略的鹈鹕算法(DEAMPOA)与加权熵法联合优化的模糊C-均值聚类(FCM)信息熵点云简化算法。采用动态自适应种群混合策略,同时融合了精英反向化思路,显著提升了鹈鹕优化算法(POA)的收敛趋势和全局寻优能力,提高了寻找FCM最优聚类中心的成功率;利用DEAMPOA结合加权熵法对FCM进行优化,提高鲁棒性的同时增强了搜索精度,得到较好的聚类结果;在8种UCI标准数据集上与4种算法对比进行聚类性能评估实验,验证了所提方法综合性能优越;将所提方法与信息熵融合,并应用在三维点云KITTI数据集简化中。实验结果表明:与包围框简化法、随机采样简化法和特征选择简化法对比,所提方法全局误差简化前后点集之间平均欧式距离(MED)指标分别降低了2.25%、6.93%、5.74%,点云简化效果最优且运行速度满足要求。

基于AquaCrop模型的水稻多目标灌溉制度优化研究

【目的】优化现有水稻灌溉制度,节约灌溉用水量,减少稻田氮磷流失量以降低面源污染风险。【方法】构建了基于AquaCrop模型和NSGA-Ⅱ算法的水稻灌溉制度模拟优化模型,利用水稻田间试验数据,开展以产量最大、氮磷流失量最小、灌水次数最少的3种目标组合(产量-氮磷流失量(Y-TNP)、产量-灌水次数(Y-N)、产量-氮磷流失量-灌水次数(Y-TNP-N))下的灌溉制度优化模型研究,提出适应不同生育期降水年型的稳产-控污-提效灌溉制度。【结果】(1)与常规灌溉相比,Y-TNP灌溉制度优化水稻产量下降2.14%,氮磷流失量减少23.09%;Y-N灌溉制度优化水稻产量下降1.76%,灌水次数减少53%;Y-TNP-N灌溉制度优化水稻产量下降2.64%,氮磷流失量减少22.83%,灌水次数减少2次。(2)不同典型年以Y-TNP-N为目标优化的稳产-控污-提效灌溉制度水稻产量介于7.74~7.78 t/hm^(2),同时大幅度减少灌水量进而降低氮磷流失量。【结论】AquaCrop模型可模拟试验区水稻的生长发育过程,本文构建的模拟-优化耦合模型可用于优化不同生育期降水年型下稳产-控污-提效的灌溉制度。

基于点-轴理论的东北亚旅游空间结构优化研究

东北亚涵盖朝鲜、韩国、日本、蒙古国4个国家全境和中国的东北3省与内蒙古自治区及俄罗斯的远东与西伯利亚2个联邦管理区,是世界经济的重要板块。本文根据点—轴理论从旅游资源、城市布局、自然地理环境、人口规模、交通线路与地域关系等角度探析了东北亚旅游空间结构的优化布局。结果显示:东北亚旅游空间结构优化形成“点—线—面”的布局,构建9核26心78点的三级旅游节点与4纵2横的旅游发展主轴线以及6大旅游发展组团的空间布局;东北亚旅游空间结构优化的轴线分为主轴线、辅助线和外联轴线,其中主轴线包括4纵2横;东北亚旅游空间可进一步划分为6大组团;东北亚旅游空间结构优化有利于区域经济健康持续发展。

改进势场蚁群算法优化路径自动规划

针对势场与蚁群联合的自主路径规划存在局部最优解、初始路径选择随机导致效率不高及环境适应性差等问题,提出了基于自适应域和参数自适应设置的改进算法。算法首先基于自适应域改进势场目标不可达问题,并过滤震荡点以平滑路径;其次,通过状态转移函数和信息素更新等相关参数的自适应设置,提高算法在收敛效率和搜索能力上的平衡性,进而提高对障碍环境的适应性。实验结果表明,所提算法能够有效避免局部最优、目标不可达和复杂环境的适应性问题,在路径长度和效率上优于实验采用的已有算法,从而验证了算法的有效性。

基于改进黏菌算法的光伏多峰值MPPT控制

针对传统最大功率点跟踪技术在局部遮阴等天气条件下存在无法追踪到全局最大功率点的问题,提出一种基于改进黏菌算法的MPPT控制。首先,对太阳电池模型及多峰值特性进行分析;其次,在黏菌算法中引入领导者策略和基于最优个体的凸透镜反向学习策略,在提高算法计算精度、收敛速度的同时克服了算法易“早熟”现象;最后,根据光伏阵列最大功率输出特性分别确定算法优化模型、初始化位置及重启机制。仿真与实验结果表明:基于改进黏菌算法的MPPT控制能快速、准确地跟踪到全局最大功率点,有效规避陷入局部最优问题,提高了光伏系统的转换效率。

含SOP的交直流混合配电网日前优化调度

为实现交直流混合配电网的高效运行,提出一种含智能软开关的交直流配电网优化调度方法。在对应用于交直流配电网的智能软开关工作原理进行阐述的基础上,建立含智能软开关的交直流配电网优化调度模型。通过线性化和凸松弛技术,将所建立的非线性优化模型转化为二阶锥规划模型,并且采用改进的50节点算例分析验证模型的有效性。算例结果表明,基于所建模型得到的智能软开关运行策略能够降低配电网运行损耗及改善电压越限的情况,显著提高混合配电网的经济性。

低烈度区某超高层塔楼优化分析

超高层建筑由于单塔体量较大,结构经济指标备受关注,寻找最合理结构体系是超高层建筑不断自我迭代更新的主题。高烈度区和低烈度区塔楼控制方向完全不同,着重介绍优化过程以寻找低烈度区超高层塔楼的最优结构。以某位于低烈度6度区的建筑高度250m超高层塔楼为例,根据塔楼初步方案确定控制指标为刚重比和剪重比,以控制指标为目标进行附加质量和自重的迭代优化。其次是提高结构刚度,核心筒尺寸根据竖向交通及设备管井已确定,在建筑允许条件下通过柱的不同排布实现了更好的框架刚度贡献;纯混凝土柱重量贡献大于刚度贡献,通过对比叠合柱和型钢柱,选取了更合理的叠合柱截面形式;在低烈度区结构竖向承载满足基本需求而水平刚度不满足规范要求的情况下,通过增加伸臂桁架提高整体刚度;通过分析不同伸臂桁架形式,选取了最有效的伸臂形式。通过降低结构冗余重量和有效提高结构刚度选出最高效结构,且建筑品质优、造价合理。

基于改进PSO算法的光伏阵列MPPT研究

为解决传统粒子群优化(PSO)算法在寻优过程中出现粒子早熟、收敛速度慢、易陷入局部优化等问题,提出一种基于反向学习的Logistic-Tent双重混沌映射和时变双重压缩因子(TVCF)策略的改进粒子群优化(LT-TVCFPSO)算法,在传统PSO算法基础上,引入了Logistic-Tent混沌映射和TVCF,既可增强种群多样性,避免粒子早熟,跳出局部优化,又能加快粒子收敛,提升全局寻优能力。最后在MATLAB/Simu-link上进行仿真。仿真结果表明:相比于传统MPPT算法,LT-TVCFPSO算法能够快速准确地追踪到全局最大功率点(GMPP)。