Effect of Quadrat Shape on Spatial Point Pattern Performance of Haloxylon ammodendron

In this study, we investigated the natural growth of Haloxylon ammodendron forest in Moso Bay, southwest of Gurbantunggut Desert. Random sample analysis was used to analyze the spatial point pattern performance of Haloxylon ammodendron population. ArcGIS software was used to summarize and analyze the spatial point pattern response of Haloxylon ammodendron population. The results showed that: 1) There were significant differences in the performance of point pattern analysis among different random quadrants. The paired t-test for variance mean ratio showed that the P values were 0.048, 0.004 and 0.301 respectively, indicating that the influence of quadrat shape on the performance of point pattern analysis was significant under the condition of the same optimal quadrat area. 2) The comparative analysis of square shapes shows that circular square is the best, square and regular hexagonal square are the second, and there is no significant difference between square and regular hexagonal square. 3) The number of samples plays a decisive role in spatial point pattern analysis. Insufficient sample size will lead to unstable results. With the increase of the number of samples to more than 120, the V value and P value curves will eventually stabilize. That is, stable spatial point pattern analysis results are closely related to the increase of the number of samples in random sample square analysis.

基于改进PointPillars的3D目标检测算法

针对自动驾驶场景下远距离及遮挡目标识别问题,对PointPillars算法进行了改进。引入了并行的空间注意力和通道注意力机制,增强了目标的位置信息及有用特征通道权重,提高了远距离目标的检测精度。在2D CNN骨干网络中引入了自适应空间特征融合模块,解决了特征拼接的信息丢失问题,提高了遮挡目标的检测精度。基于KITTI数据集在3种不同场景难度下分别对SECOND、PointPillars、改进PointPillars这3种算法进行了定量分析验证,并将改进的PointPillars算法进行可视化分析。定量分析表明,改进PointPillars算法在鸟瞰图模式下目标检测精度最大提升2.75%;在三维模式下目标检测精度最大提升2.93%;在AOS模式下目标检测精度最大提升4.05%,可视化结果表明改进的PointPillars算法能有效检测远距离及遮挡目标。

Spatial patterns of Picea crassifolia driven by environmental heterogeneity and intraspecifi c interactions

Research on the spatial patterns of tree populations is critical for understanding the structure and dynamic processes of forests.However,little is known about how the underlying drivers shape these patterns and species interactions in forest systems.In this study,spatial point pattern analysis investigated the combined eff ects of intraspecifi c interactions and environmental heterogeneity on the spatial structure and internal maintenance mechanisms of Picea crassifolia in the Qilian Mountain National Nature Reserve,China.Data were obtained from a 10.2-ha dynamic monitoring plot(DMP)and sixteen 0.04-ha elevation gradient plots(EGPs).Under complete spatial randomness,both mature trees and saplings in the DMP demonstratedlarge-scale aggregation with negative correlations.In EGPs,saplings were clustered in small mesoscales,mature trees were randomly distributed,and the interactions of saplingstrees at all elevations were not correlated.By eliminating the interference of environmental heterogeneity through the inhomogeneous Poisson process,saplings in the DMP and EGPs were clustered in small scales and trees randomly distributed.Intraspecifi c associations were negatively correlated,in the DMP and at low elevations,and no correlations in high elevations of EGPs.In the vertical scale,saplings showed a small-scale aggregation pattern with increase in elevation,and the aggregation degree fi rst decreased and then increased.The interactions of saplings-trees and saplings–saplings showed inhibitions at small scales,with the degree of inhibition gradually decreasing.Spatial patterns and associations of adults–adults did not change signifi-cantly.The results revealed that intraspecifi c interactions and environmental heterogeneity regulated the spatial patterns of P.crassifolia at small and large scales,respectively.Environmental heterogeneity might be the most decisive factor aff ecting the spatial patterns of saplings,while trees were more aff ected by intraspecifi c interactions.Moreover,competition between trees in this area could be more common than facilitation for the growth and development of individuals.

Formula for calculating spatial similarity degrees between point clouds on multi-scale maps taking map scale change as the only independent variable

The degree of spatial similarity plays an important role in map generalization, yet there has been no quantitative research into it. To fill this gap, this study first defines map scale change and spatial similarity degree/relation in multi-scale map spaces and then proposes a model for calculating the degree of spatial similarity between a point cloud at one scale and its gener- alized counterpart at another scale. After validation, the new model features 16 points with map scale change as the x coordinate and the degree of spatial similarity as the y coordinate. Finally, using an application for curve fitting, the model achieves an empirical formula that can calculate the degree of spatial similarity using map scale change as the sole independent variable, and vice versa. This formula can be used to automate algorithms for point feature generalization and to determine when to terminate them during the generalization.

PointPCA:一种三维空间点云的特征提取算法

点云是一个庞大点的集合而且拥有重要的几何结构。由于其庞大的数据量,不可避免地就会在某些区域内出现一些相似点,这就使得在进行特征提取时提取到一些重复的信息,造成计算冗余,降低训练的准确率。针对上述问题,提出了一种新的神经网络——PointPCA,可以有效地解决上述问题;在PointPCA中,总共分为三个模块:a)采样模块,提出了一种average point sampling(APS)采样方法,可以有效地规避一些相似的点,得到一组近似代表这组点云的新的点集;b)特征提取模块,采用分组中的思想,对这组新的点的集合进行多尺度空间特征提取;c)拼接模块,将每一尺度提取的特征向量拼接到一起组合为一个特征向量。经过实验表明,PointPCA比PointNet在准确率方面提升了4.6%,比PointNet++提升了1.1%;而且在mIoU评估测试中也有不错的效果。

Spatial Pattern and Influencing Factor Analysis of Attended Collection and Delivery Points in Changsha City, China

Attended collection and delivery points are vital components of ‘last-mile logistics’.Based on point of interest(POI) data for Cainiao Stations and China Post stations in Changsha City, China, this paper provides a detailed exploration of the basic features, spatial distribution, and location influencing factors of attended collection and delivery points.Specifically, analyses of the types, service objects and location distributions of the attended collection and delivery points alongside a discussion of their spatial pattern and influencing factors provides a reference for their general geographic layout and characteristics.The findings of this study indicate that: 1) The main mode of operation of attended collection and delivery points is franchises, with other modes of operation rely on supermarkets and other individual shop types.2) The main service targets of attended collection and delivery points are communities, schools, and businesses, followed by townships, enterprises, scenic spots, and administrative units.3) Approximately 77.44% of the attended collection and delivery points are located near the exits of service areas;others are situated in the centre of the service areas.For the Cainiao Stations, 80% are located within 125 m of the exit;for the China Post stations, 80% are located within 175 m of the exit.4) The spatial distribution of the attended collection and delivery points in Changsha is unbalanced, with ‘more centre and fewer surrounding’.The centre is an ‘inverted triangle’, and the edge is an ‘orphan’, showing a northwest-southeast orientation and symmetrical along the axis.The layout of the attended collection and delivery points forms three core areas, and the number of sites decreases with the distance from the core.5) The number and distribution of the attended collection and delivery points are strongly consistent with the regional economic development level, population, and roadway system traffic convenience.Most attended collection and delivery points are on residential, scientific and educational, and commercial and financial land.

The Spatial Patterns of Road Traffic Crash Black Spots and Emergency Facilities in Federal Capital City

This study aimed at determining the spatial patterns of Road Traffic Crash (RTC) black spots, Federal Road Safety Commission (FRSC) zebra points and emergency health care facilities in Federal Capital City (FCC). The aim was to provide stakeholders with information that will aid their understanding of accident prone locations and accessible rescue possibilities for accident victims on the roads in FCT. GPS Map 76S Mark (GARMIN) was used to locate and pick coordinates of data in the study area. A total of 16 possible emergency health care facilities, seventy (70) RTC black spots and Five Zebra point locations were obtained from FRSC. ArcGIS 10.0 was used to compute the data by plotting the coordinates to produce maps of the spatial relationship and to carry out Nearest Neighbour Analysis (NNA). The result was further used to determine the spatial patterns of RTC black spots as well as patterns of the emergency facilities. Generally, the result shows that the spatial trend is turning towards dispersion. However, there is less than 1% likelihood that the dispersed patterns could be the result of random chance. It was recommended that, the Federal Road Safety Commission should be staffed with trained professionals that can be responsible for accident data surveillance and analysis using geospatial techniques.

基于空间传播的多视图三维重建

针对多视图三维重建任务中点云完整性欠佳的问题,提出一种基于空间传播的多视图深度估计网络(SPMVSNet)。引入空间传播思想用于复杂条件下的稠密点云重建,并分别设计基于空间传播的混合深度假设策略和空间感知优化模块。混合深度假设策略采用由粗糙到精细的深度推理方式,将深度估计视为多标签分类任务,对正则化概率体执行交叉熵损失以约束代价体,从而避免回归方法过拟合和收敛速度过慢的问题。空间感知优化模块从包含高级语义特征表示的特征图中获得引导,在进行置信度检查后采用卷积空间传播网络,通过构建亲和矩阵来细化最终的深度图。同时,为解决大多数方法存在的对不满足多视图一致性的不可靠区域重建质量较低的问题,进一步结合注意力机制设计具有样本自适应能力的动态特征提取网络,用于增强模型的局部感知能力。实验结果表明,在DTU数据集上,SP-MVSNet的重建完整性相比于CVP-MVSNet提升32.8%,整体质量提升11.4%。在Tanks and Temples基准和Blended MVS数据集上,SP-MVSNet的表现也优于大多数已知方法,取得了良好的三维重建效果。

基于手机信令数据的城市居民动态OD矩阵提取方法

现有的城市居民出行调查周期较长,交通小区划分粒度粗糙,导致调查不能及时准确地获取居民出行信息。针对该问题,提出了一种基于手机信令数据的城市居民动态OD矩阵提取方法。首先,针对信令数据中的两种复杂噪声:乒乓切换和漂移数据,提出了基于窗口阈值的检测与等效位置替换方法,以及复杂漂移点的检测和标记处理方法;然后,提出一种改进的ST-DBSCAN聚类方法,引入一种等时化方法将时间信息与空间信息相结合,识别出行过程中的驻留点;最后,基于地理信息系统构建含有道路关键节点的路网,将居民出行OD与路网节点相匹配,有效推导出城市居民动态OD矩阵。实验结果表明:与ST-DBSCAN算法相比,所提改进的ST-DBSCAN算法在聚类效果和识别速度上分别提升了6.10%和5.26%;与统计方法和二阶统计量方法相比,基于改进的ST-DBSCAN算法的动态OD矩阵提取方法在均方误差(MSE)上分别降低了16.98%和21.55%。以北京市为例,运用提出的动态OD矩阵提取方法,能够及时有效地分析城市居民日常与高峰时段的出行特征。

Fractal Analysis on the Spatial Structure of Land Use Patterns in a Non-Point Source Polluted Area in Southern China

Based on the TM remote sensing interpretation of land use data of a non-point source polluted area in southern China in 2015, supported by ArcGIS and the fractal theory, this paper established the perimeter-area fractal equation of the spatial structure of land use types, quantitatively calculated the fractal dimension and stability index of various land use types in this non-point source polluted area in 2015, and analyzed the spatial structure characteristics of land use types. The results showed that the fractal dimension values of the different land use types in the research area ranged 1.141–1.657, and the whole fractal dimension of the research area was 1.206. To be specific, the fractal dimension values of unused land, grass land and rural residential land were high, and the fractal dimension values of dry land, paddy filed and waters were low, and the fractal dimension values of urban land, industrial traffic construction land and forest land were moderate. The area of the land use types whose stability index values were 0.2–0.4 accounted for 89.5% of total area, suggesting that the stability of the spatial structure of land use types of Beibei District in 2015 was not high generally. The rural residential land and urban land of the research area should be planned and managed more properly from now on.

煤巷掘进过程中煤岩破裂微震事件研究

目的为了监测和预警煤矿掘进巷道前方可能发生的危险,方法利用微震监测技术实时监测,利用监测系统捕获到的数据信息研究掘进影响下微震事件的空间分布规律、发生时间、频次和能量关键点,然后利用短时傅里叶变换和S变换方法分析典型煤岩破裂微震信号的时频特征。结果结果表明:(1)微震事件主要分布在沿掘进巷道前方50~100 m区域,大多为中低能量微震事件,仅在掘进工作面前方出现少量强度较大的煤岩破裂事件,且震级多为-2~0;(2)掘进速度和发生微震事件的数量及能量大小成正相关,同时掘进速度也影响着煤岩动力灾害的发生;(3)巷道掘进会造成煤岩体失稳,其中岩体破裂信号频率高,为0~2000 Hz,比重值大,主频约700 Hz;煤岩破裂微震事件信号频率为90~1000 Hz,主频约250 Hz;岩体破裂后,支撑力下降,压力转移给煤体,导致煤体破裂,煤体破裂微震事件信号频率为90~1000 Hz,主频约350 Hz。结论微震事件能够综合反映煤岩体受力破坏程度,判别掘进过程中煤岩破裂类型,为矿井生产中煤岩动力灾害的预警提供理论基础。

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。

秦岭南麓小流域乡村聚落空间集聚PSR评价研究——以乾佑河流域柞水段为例

小流域是山地区域乡村聚落广泛存在的空间载体,随着新型城镇化建设、乡村振兴战略的实施,这里的聚落也开始发生剧烈变动,开始了更为集约的收缩型集聚。为了更好地发现其变化集聚的状态与趋势,对聚落发展提出合理的优化建议与引导,急需合适的评价工具与方法。因此,针对秦岭南麓小流域内聚落斑块零散存在的现实情况,借助“压力-状态-响应”评价模型,提出了聚焦小流域乡村聚落社会-生态健康发展的PSR评价方法以及聚焦关键问题的研究思路;进而以乾佑河柞水段的七条典型小流域为对象,借助核密度估算工具,进行内部聚落空间集聚特征的总结,建构起可进行比较的核点集聚区域、点轴集中区概念;并通过评价参数的筛选,明确了小流域、核点集聚区域双层级PSR评价体系,结合实践提出相关优化建议,也为小流域乡村聚落合理建设探索更科学的研究基础。

Spatial and Temporal Distribution Characteristics of Agricultural Non-point Source Pollution in Xixi Watershed of Jinjiang Basin

The SWAT model was applied to analyze the temporal-spatial distribution patterns of non-point source pollution loads and the difference of pollution loads of different land use types in Xixi Watershed of Jinjiang Basin. The results showed that both yearly nitrogen and phosphorus pollution loads were evenly distributed during 1973 to 1979,the annual TN pollution from non-point source was 1530 t,or 6. 3 kg / ha,and the annual TP pollution from non-point source was 270 t,or 1. 1 kg / ha during 1973 to 1979 in the watershed. Considerable differences were identified on both monthly nitrogen and phosphorus pollution loads. The TN and TP pollution loads during the flood season( from April to September) accounted for 76. 2% and 75. 8% of the annual load respectively. There were great differences in both TN and TP pollution loads of different land use types in the study area,and the pollution load of both farmland and orchard was higher than that of the other land use types. TN and TP pollution loads of farmland accounted for 66% and 83% of total watershed. There was a great spatial difference in the nonpoint source pollution load of the study area. The critical source areas of non-point source pollution are mainly located at Guanqiao Town,Longmen Town,Changkeng Town,Shangqing Town and Dapu Town,where the efforts of controlling pollution should be made.

基于Voronoi图的空间点事件统计聚类方法

挖掘地理空间数据中点事件聚集模式对于揭示流行疾病、犯罪分布热点区域及城市基础设施空间分布格局等具有重要意义。针对不同形状、密度和大小的显著空间点聚集模式的识别,目前以空间扫描统计为代表的方法虽然可以对空间点聚类的显著性进行统计推断,减少虚假聚类结果,但其主要用于识别球形或椭圆形状的聚簇,对于沿着街道或河道分布的任意形状、不同密度的显著空间点聚簇识别还存在局限。因此,本研究提出一种基于Voronoi图的空间点聚集模式统计挖掘方法。首先,采用Voronoi图来度量空间点分布的聚集性,将空间点聚类问题转化为热点区域探测问题;其次,结合局部Gi*统计量探测统计上显著的空间点聚簇;最后,通过模拟数据和真实犯罪事件数据进行实验与对比分析。结果表明:本方法能够有效探测任意形状的空间点聚类,并对空间点簇的显著性进行统计判别,识别显著的空间点簇,减少随机噪声点的干扰;聚类识别结果优于现有代表性方法,如DBSCAN算法、空间扫描统计方法等。

城市休闲产业聚类模式APM算法模型开发与校验

城市休闲相关产业的高质量发展对当前我国城市消费升级以及人居环境质量提升具有重要现实意义。但是,现有研究未能精准地捕捉海量广域分布的城市休闲产业的基本空间分布规律与结构,而已有的空间聚类算法较多适用于城市用地分析,未能很好地适用于离散分布的城市休闲产业研究。为此,文章基于空间兴趣点数据,开发距离通达值及空间集群中心点等算法,构建城市休闲旅游产业聚类模式空间算法模型(APM)。在以广州为例的研究中,APM模型捕捉出3170个以500 m步行生活圈为范围的城市休闲产业集群,校验了APM模型的科学性与应用价值。整体上,APM算法可以较好地捕捉城市休闲业态集群的空间结构,清晰识别城市休闲产业空间冷、热点分布的基本结构,由其捕捉行程的聚类边界与实际道路和建筑走向、水系边界、区域范围等重合度高,聚类集群符合实际情况,具备可信度与有效性。该研究是休闲产业集聚机制研究的一次方法创新,在算法精度、实际应用、可视化效率上均做出了创新性推进。与Fishnet方法相比,可以更科学精准地识别城市内部多个休闲消费商圈的边界,实现了高效率的城市休闲产业集群捕捉;与同位模型相比,可以呈现多类别的城市休闲业态结构,突破了现有研究只能捕捉两类业态组团的局限。

基于移动多线激光雷达扫描的树冠叶面积估计方法

移动单线激光雷达(Laser detection and ranging,LiDAR)扫描(Mobile single-layer LiDAR scanning,MSLS)树冠叶面积估计方法使用单一视角的单线激光雷达采集树冠点云数据,获取的冠层信息不够全面,限制了树冠叶面积估计精度。本文提出一种基于移动多线LiDAR扫描(Mobile multi-layer LiDAR scanning,MMLS)的树冠叶面积估计方法,使用多线LiDAR从多个视角采集树冠点云数据,提升树冠叶面积估计精度。首先,将多线LiDAR采集的点云数据变换到世界坐标系下,通过感兴趣区域(Region of interest,ROI)提取出树冠点云。然后,提出一种MMLS树冠点云融合方法,逐个融合单个激光器采集的树冠点云,设置距离阈值删除重复点,添加新点。最后,构建MMLS空间分辨率网格,建立基于树冠网格面积的树冠叶面积估计模型。实验使用VLP-16型多线LiDAR传感器搭建MMLS系统,设置1、1.5 m 2个测量距离和间隔45°的8个测量角度对6个具有不同冠层密度的树冠进行数据采集,共得到96个树冠样本。采用本文方法,树冠叶面积线性估计模型的均方根误差(Root mean squared error,RMSE)为0.1041 m^(2),比MSLS模型降低0.0578 m^(2),决定系数R^(2)为0.9526,比MSLS模型提高0.0675。实验结果表明,本文方法通过多线LiDAR多视角树冠点云数据采集、MMLS树冠点云融合和空间分辨率网格构建,有效提升了树冠叶面积估计精度。

协同随机森林方法和无人机LiDAR空谱数据的盐沼植被“精灵圈”识别

“精灵圈”作为一种典型的空间自组织结构,对盐沼植被生态系统及其功能有重要影响。获取“精灵圈”的空间格局及时空变化,可为厘清其生态演化机理提供重要科学支撑。本文基于随机森林机器学习方法,结合无人机激光雷达(LiDAR)点云的空间信息与光谱信息,对盐沼植被“精灵圈”进行智能识别与提取。首先,利用激光雷达方程和Phong模型,消除距离、入射角以及镜面反射效应对强度数据的影响,并且通过校正后强度数据滤波分离植被点云与地面点云。然后,构造系列空间特征及几何变量,利用随机森林算法,对植被点云中的正常植被和“精灵圈”进行分类。结果表明:该方法无需人工经验设置参数,能够精确地从无人机LiDAR三维点云数据中快速自动识别“精灵圈”,总体精度为83.9%。本文为“精灵圈”时空分布反演提供了一种高精度的方法,也为基于机器学习的三维点云数据处理提供了技术借鉴。

林窗对黄檗幼树空间分布及生长发育的影响

【目的】林窗增加了林分空间结构的复杂性和林下生境的异质性,对林下幼树的更新具有重要影响。黄檗是我国国家重点保护野生植物,研究黄檗幼树的生长状况、空间分布与林窗空间结构的关系,有助于了解其对林下生境的适应性,对促进黄檗种群的更新具有重要意义。【方法】本文以2014年在北京百花山自然保护区内林下栽种的黄檗幼树为研究对象,于2021年对黄檗的生长及存活状况进行了调查,并使用背包式激光雷达获取了样地的林窗空间结构数据。采用空间点格局、生境关联性检验和方差分析等方法研究了林窗对黄檗幼树空间分布及生长状况的影响。【结果】(1)自然生长条件下,7年后黄檗幼树存活率为19.25%,在排除扩散限制后,存活黄檗幼树的空间分布表现为聚集分布。(2)林窗空间结构是黄檗幼树空间分布格局形成的重要原因,黄檗幼树分布与林窗边缘呈显著正关联,而与林冠区呈显著负关联。(3)黄檗幼树的基径在林窗边缘显著高于林冠区和林窗中心(P<0.01),在林冠区与林窗中心之间没有显著差异;黄檗幼树的冠幅和树高则呈现显著的梯度变化(P<0.05),即林窗边缘>林冠区>林窗中心。【结论】研究表明林窗对黄檗幼树的空间分布格局和生长状况具有重要的影响。林窗边缘的生境条件更有利于黄檗幼树的存活和生长,而林窗中心和林冠区则会对黄檗幼树的生长发育产生一定的抑制作用。

基于多传感融合的巷道三维空间映射

针对我国煤矿掘进工作面智能感知能力不足的问题,提出了一种基于多传感融合的煤矿巷道三维空间映射方法。该方法利用激光雷达等3D传感器预览煤矿巷道环境信息,并构建实时自主映射模型,为大型煤机装备提供空间感知信息。基于姿态感知理论的多传感融合技术,对扩展卡尔曼滤波器算法进行优化,通过局部误差状态滤波估计和全局状态迭代估计,提出了一种基于距离残差的空间映射方法,实现了在几何退化的巷道环境下全局地图的快速更新和映射的高精度。其主要贡献在于将多传感器数据进行融合,构建出高精度的地下三维地图,并提高了巷道环境的智能感知能力和自主操作水平。首先,采用先进的激光雷达(LiDAR)传感器获取巷道内部的点云数据,提取点云特征,并建立帧间特征映射模型,从而构建出厘米级别的地下三维地图。同时,通过迭代误差状态的卡尔曼滤波器实现激光雷达与惯性导航的数据融合,提高了系统的鲁棒性,确保了地下三维地图的准确性和稳定性。此外,还考虑了井下粉尘和水雾等因素对巷道三维空间映射精度的影响,优化卡尔曼滤波增益调节激光雷达里程计权重,增强了低质量浓度粉尘条件下三维空间映射的环境适应性。实验结果表明,相较于LeGO-LOAM和LINS,笔者提出的三维空间映射方法在轨迹估计准确度、映射地图的点云数量以及点云密度等方面均表现更优。笔者所提出的方法可用于高精度、快速更新三维地图的自主掘进装备,提高煤矿掘进工作面的智能化水平,为巷道环境的智能化感知和自主作业提供了数据支持。