Fuzzy topological relations between fuzzy spatial objects

Fuzziness is an internal property of spatial objects.How to model fuzziness of a spatial object is a main task of next generation GIS.This paper proposes basic fuzzy spatial object types based on fuzzy topology.These object types are the natural extension of current nonfuzzy spatial object types.A fuzzy cell complex structure is defined for modeling fuzzy regions,lines and points.Furthermore,fuzzy topological relations between these fuzzy spatial objects are formalized based on the 9intersection approach.This model can be implemented for GIS applications due to its scientific theory basis.

Spatial Object Aggregation Based on Data Structure, Local Triangulation and Hierarchical Analyzing Method

This paper focuses on the methods and process of spatial aggregation based on semantic and geometric characteristics of spatial objects and relations among the objects with the help of spatial data structure (Formal Data Structure),the Local Constrained Delaunay Triangulations and semantic hierarchy.The adjacent relation among connected objects and unconnected objects has been studied through constrained triangle as elementary processing unit in aggregation operation.The hierarchical semantic analytical matrix is given for analyzing the similarity between objects types and between objects.Several different cases of aggregation have been presented in this paper.

SPATIAL OBJECTS AND SPATIAL RELATIONSHIPS

Spatial object and spatial relationship are two basic concepts of GIS.Spatial object is the digital representation of geographical entity or phenomenon,which forms the basis for data management and analysis;spatial relationship is theconnexion between spatial objects when geometric properties are considered.Thecommonly used classification of spatial objects as points,lines and areas is mathe-matically strict,and suitable for data management,but a bit too generalized forrepresenting real entities and extracting spatial relationships.A good classificationmodel should not only be good for representing real entities,but also good for re-vealing spatial relationships,therefore good for formalizing spatial analyses.

DEALING WITH SPATIAL RELATIONS FOR SPATIAL OBJECTS WITH RANDOMNESS IN GIS

To design retrieval algorithm of spatial relations for spatial objects with randomness in GIS,this paper builds up the membership functions based on set theory idea,used for determination of topological spatial relations between random objects,such as between point and point,point and line or polygon,which provides theoretical basis for retrieving spatial relations between certain and random objects.Finally,this paper interprets detailed methods and steps of realizing them by means of some simple examples under the GIS’s environment.

Behavior Mining of Spatial Objects with Data Field

The advanced data mining technologies and the large quantities of remotely sensed Imagery provide a data mining opportunity with high potential for useful results. Extracting interesting patterns and rules from data sets composed of images and associated ground data can be of importance in object identification, community planning, resource discovery and other areas. In this paper, a data field is presented to express the observed spatial objects and conduct behavior mining on them. First, most of the important aspects are discussed on behavior mining and its implications for the future of data mining. Furthermore, an ideal framework of the behavior mining system is proposed in the network environment. Second, the model of behavior mining is given on the observed spatial objects, including the objects described by the first feature data field and the main feature data field by means of the potential function. Finally, a case study about object identification in public is given and analyzed. The experimental results show that the new model is feasible in behavior mining.

Ontological exploration of geospatial objects in context

Structured study of spatial objects and their relationships leads to a better cognition of the geospatial information and creates the concept of context at a higher level of abstraction.This study is aimed at providing a comprehensive definition of the context for geospatial objects.A combination of binary qualitative spatial relationships(i.e.direction,distance,and topological relations)among the members of a set of spatial objects will be used accordingly.In addition,by incorporating the general concept of context,obtained from either static data(attributes in a database)or dynamic data(sensors),the compact context of spatial objects will be introduced.Our framework for presentation of the involved knowledge and conception about the objects in context is also explored using ontology and description logic because of powerful conceptualization of relationships,either spatial or non-spatial,integrally.For this purpose,the hierarchies of main structure and object properties are formed at first.The constraint and characteristics of classes,such as subclasses,equivalent classes,cardinality etc.,and object properties,such as being functional,transitive,symmetric,asymmetric,inverse functional,disjoint etc.,are discovered and presented in more detail using web ontology language in description logic mode.The implementation is then performed in the framework of semantic web and extensible markup language syntaxes.The method ultimately facilitates,spatial reasoning by effective querying in a semantic framework taking pellet reasoner and SPARQL(a recursive acronym for SPARQL Protocol and RDF Query Language).

OBJECT-ORIENTED SPATIAL DATA ORGANIZATION AND MANAGEMENT IN GEOSTAR

Spatial data, including geometrical data, attribute data, image data and DEM data, are huge in volume and relations among them are complex. How to effectively organize and manage those data is an important problem in GIS. Several problems about space data organization and management in GeoStar which is a basic GIS software made in China are discussed in this paper. The paper emphasizes on object model of spatial vector, data organization, data management and how to realize the goal, and the like.

Object detection based on combination of local and spatial information

A method of object detection based on combination of local and spatial information is proposed. Firstly, the categorygiven representative images are chosen through clustering to be templates, and the local and spatial information of template are ex- tracted and generalized as the template feature. At the same time, the codebook dictionary of local contour is also built up. Secondly, based on the codebook dictionary, sliding-window mechanism and the vote algorithm are used to select initial candidate object win- dows. Lastly, the final object windows are got from initial candidate windows based on local and spatial structure feature matching. Experimental results demonstrate that the proposed approach is able to consistently identify and accurately detect the objects with better performance than the existing methods.

An Effective Image Retrieval Mechanism Using Family-based Spatial Consistency Filtration with Object Region

How to construct an appropriate spatial consistent measurement is the key to improving image retrieval performance. To address this problem, this paper introduces a novel image retrieval mechanism based on the family filtration in object region. First, we supply an object region by selecting a rectangle in a query image such that system returns a ranked list of images that contain the same object, retrieved from the corpus based on 100 images, as a result of the first rank. To further improve retrieval performance, we add an efficient spatial consistency stage, which is named family-based spatial consistency filtration, to re-rank the results returned by the first rank. We elaborate the performance of the retrieval system by some experiments on the dataset selected from the key frames of ＂TREC Video Retrieval Evaluation 2005 （TRECVID2005）＂. The results of experiments show that the retrieval mechanism proposed by us has vast major effect on the retrieval quality. The paper also verifies the stability of the retrieval mechanism by increasing the number of images from 100 to 2000 and realizes generalized retrieval with the object outside the dataset.

Overcoming Object Misalignment in Geo-Spatial Datasets

In integrating geo-spatial datasets, sometimes layers are unable to perfectly overlay each other. In most cases, the cause of misalignment is the cartographic variation of objects forming features in the datasets. Either this could be due to actual changes on ground, collection, or storage approaches used leading to overlapping or openings between features. In this paper, we present an alignment method that uses adjustment algorithms to update the geometry of features within a dataset or complementary adjacent datasets so that they can align to achieve perfect integration. The method identifies every unique spatial instance in datasets and their spatial points that define all their geometry;the differences are compared and used to compute the alignment parameters. This provides a uniform geo-spatial features’ alignment taking into consideration changes in the different datasets being integrated without affecting the topology and attributes.

图卷积神经网络综述

图卷积神经网络是图论与深度学习的交叉,已成为机器学习领域的研究热点。基于此,介绍了图卷积神经网络的形成,梳理了两大类经典的图卷积神经网络:谱方法和空间方法,详细介绍了这两类图卷积神经网络模型,分析了图卷积操作的核心理论基础,介绍了图卷积神经网络在各领域的应用,总结了图卷积神经网络面临的主要挑战,展望了图卷积神经网络的发展趋势,并分析了图卷积神经网络在野外环境下蝴蝶识别任务中的潜在应用。

非连通空间对象方向关系表达与推理

为了弥补现有的二维空间对象方向关系表达模型大都利用点、最小外包矩形等近似地代替空间对象,距离真实空间对象间方向关系的描述与推理仍存在差距的不足,提出了一种基于Voronoi图的非连通空间对象方向关系表达模型。该模型借助Gestalt心理学理论,通过提取非连通空间对象的特征点、特征链,构建空间对象间的可视区域,生成方向关系Voronoi图,实现了非连通、含洞的参考对象与目标对象间方向关系的表达。该模型较好地顾及了空间对象形状、大小等因素带来的影响,表达精度更高、适用范围更广。为了提高复杂空间对象方向关系复合推理的精度,基于该模型提出了一个非连通对象间主方向关系复合推理算法。该算法借助Tile-union运算和Pr运算,实现了该模型下基本主方向关系的复合推理,降低推理结果的不确定性。分析和验证的结果表明,提出的非连通空间对象方向关系模型及复合推理算法,提高了表达与推理的精度,完善和提高了对复杂空间对象方位关系的分析与处理能力。

论先秦文献中时间概念的多重隐喻

中国人自古以来就善于从生活中的事物入手,用具体普遍的事物来表达不可言说的抽象概念。在认知能力有限的条件下,中国早期先民借助空间方位、客观实体实现了对抽象时间概念的具体化表达,并逐渐内化为人们对时间的自觉认知,这一实现过程正是隐喻作为思维方式的结果。先秦文献中对时间的隐喻主要包含对动态时间推移的表达,以及对静态历法、时令、四季节气的描述,形成“时间是空间”与“时间是客观实体”两个根本性隐喻。

考虑偏好的空间文本对象多目标最短路径查询

为了解决路网中考虑偏好的空间文本对象多目标最短路径查询问题,设计了包含属性的IR树索引结构;然后利用类别约束,排序约束构建限定的有向无环图.提出了支持属性的局部贪婪算法SALG_SPG,以对最短路径查询进行处理;为进一步提高查询处理效率,提出了支持属性的全局优化算法SAGO_SPG,该算法利用阈值和缓存机制减少了大量无效、重复的计算;最后,通过与现有的FEBF算法的对比实验,分析并验证了所提算法的高效性.

一种基于SAM-MSFF网络的低照度目标检测方法

由于低照度图像具有对比度低、细节丢失严重、噪声大等缺点,现有的目标检测算法对低照度图像的检测效果不理想.为此,本文提出一种结合空间感知注意力机制和多尺度特征融合(Spatial-aware Attention Mechanism and Multi-Scale Feature Fusion,SAM-MSFF)的低照度目标检测方法 .该方法首先通过多尺度交互内存金字塔融合多尺度特征,增强低照度图像特征中的有效信息,并设置内存向量存储样本的特征,捕获样本之间的潜在关联性;然后,引入空间感知注意力机制获取特征在空间域的长距离上下文信息和局部信息,从而增强低照度图像中的目标特征,抑制背景信息和噪声的干扰;最后,利用多感受野增强模块扩张特征的感受野,对具有不同感受野的特征进行分组重加权计算,使检测网络根据输入的多尺度信息自适应地调整感受野的大小.在ExDark数据集上进行实验,本文方法的平均精度(mean Average Precision,mAP)达到77.04%,比现有的主流目标检测方法提高2.6%~14.34%.

语义重建的动态监督伪装物体检测

伪装物体检测旨在分离视觉上高度融入周围环境的物体,但是物体前景与背景存在大量相似干扰,导致分割过程中易于出现明显错误.针对上述问题,文中提出基于语义重建的动态监督伪装物体检测网络(Dynamic Supervised Camouflaged Object Detection Network with Semantic Reconstruction,DSSRNet),通过重建特征图的空间语义和引入置信度指导网络训练,实现对伪装物体的准确分割.首先,提出空间语义低秩重建机制,精细感知不同尺度下伪装物体具有区分性的语义特征.然后,生成预测置信度图,对分割过程进行动态监督,减少网络因过于自信造成的假阳性和假阴性判断.最后,提出模糊感知损失函数,对网络施加强约束,改善预测时产生的图像模糊问题.在3个具有挑战性的基准数据集上的实验表明,DSSRNet可较好地排除相似信息干扰,取得精准的分割效果.

基于孪生多注意力网络的接触网侵限轻飘物跟踪

针对接触网侵限轻飘物尺度变化大、遮挡干扰和铁路限界背景杂乱,易导致跟踪失败的问题,提出一种基于孪生多注意力网络的接触网轻飘物侵限跟踪新方法。引入三种注意力机制从更深层次提取轻飘物特征,通过空间注意力消除局部感受野限制,通过通道注意力突出轻飘物类别的通道特征,将交叉注意力聚焦于目标模板与搜索图像信息,采用空间正则化滤波器抑制背景干扰,融合各部分特征实现对侵限轻飘物的跟踪。采用OTB100数据集进行精度与准度实验,以国家重点实验室试验线采集的数据为算例进行实验,通过消融实验验证了新方法的有效性。结果表明:相比于相关滤波类SRDCF算法和深度学习类SiamRPN++、DaSiamRPN算法,新方法可获得更好的鲁棒性和准确性。

面向弱纹理空间目标的特征点匹配方法

特征点提取与匹配是遥感图像处理中关键的一环,目前成熟的算法大多面向对地成像类型的遥感图像,对于空间目标的遥感图像,没有考虑成像条件与探测平台的影响因素,特征点匹配质量较差。针对空间目标的匹配精度不高这一问题,文章提出了一种基于聚类的特征点匹配算法。首先,根据空间目标的重复弱纹理进行特征点提取与描述,再利用特征点的空间位置进行聚类,并对特征点簇进行匹配;之后将特征点的主方向减去目标整体方向,利用特征点主方向对每一个点簇进行再分组,并完成特征点匹配;最后利用最近邻次近邻比率方法和随机样本一致算法(RANSAC)剔除外点。采用该特征点匹配方法进行的模拟成像数据实验结果表明,对于空间目标图像,基于聚类的特征点匹配较直接匹配,匹配数量的提升最高可达50%,重投影误差优于1/4个像元。文章提出的这一方法使用目前通用的各种特征描述子,能够大幅度提高空间目标图像特征点匹配的数量与精度。

基于Oracle Spatial的GIS数据组织及查询

G IS的一个难点是实现空间数据和属性数据的一体化存储和管理,研究用Oracle Spatial技术来解决这个问题。在确定了采用对象—关系模型来存储G IS数据后,比较了空间数据的两种导入方式,在此基础上,选用EasyLoader工具将G IS数据导入到Oracle数据库中,再用四元树索引为其建立索引。最后,选择Java作为开发工具,采用JDBC方式与Oracle数据库相连并实现了对G IS数据的查询。

基于改进YOLOv5s算法的禁捕期长江渔船识别及应用研究

长江实行十年禁渔是长江生态环境修复的关键环节,针对禁渔期间长江非法捕捞渔船目标小、背景复杂、流动大等问题,提出了一种基于改进YOLOv5s的目标检测算法。该算法优化多尺度自适应锚框模块,采用改进的K-means++聚类算法,重新匹配适合长江船舶尺寸的锚框;使用轻量高效的坐标注意力(coordinate attention,CA)机制,提升模型关注目标通道信息特征的能力;采用SPPCSPPC(spatial pyramid pooling and context-aware spatial pyramid pooling combination)对特征图进行池化,提高小目标检测能力;通过构建长江船舶数据集训练得到最优权值模型。结果显示,改进后的模型在准确率、召回率、mAP0.5、mAP0.5∶0.9和原模型相比分别提高了1.5%、3.0%、2.4%、7.7%,且训练过程损失收敛更快,收敛值更低,能够准确快速识别出长江上的渔船目标。研究结果可为长江十年禁渔提供技术支持。