视觉同步定位与建图(simultaneous localization and mapping,SLAM)是实现移动机器人自主定位并构建环境地图的关键环节。SLAM技术虽能精确重建环境几何结构,却难以为机器人提供执行复杂任务所需的语义理解能力;建筑信息模型(building i...视觉同步定位与建图(simultaneous localization and mapping,SLAM)是实现移动机器人自主定位并构建环境地图的关键环节。SLAM技术虽能精确重建环境几何结构,却难以为机器人提供执行复杂任务所需的语义理解能力;建筑信息模型(building information model,BIM)包含丰富的建筑信息,但与机器人操作系统(robot operating system,ROS)之间存在显著的数据格式和表达方式差异,且现有研究多采用人工方式进行转换,效率低下难以规模化应用,且室内环境并非静态不变,从而会影响机器人的导航决策。因此,提出一种集成BIM数据的ROS室内语义地图构建与动态更新方法。通过研发工业基础类(industry foundation classes,IFC)到统一机器人描述格式(unified robot description format,URDF)自动转换器,实现从BIM到机器人仿真环境的自动化建模;融合YOLOv8与随机采样一致性(random sample consensus,RANSAC)算法,建立视觉驱动的语义地图动态更新机制。结果表明,静态建筑元素还原准确率达98%以上,动态物体识别精度达0.9以上,显著提升了语义地图的自动化程度、知识丰富度及环境适应性。展开更多
The Liupan Mountains is located in the southern Ningxia Hui Autonomous Region of China, which forms an important dividing line between landforms and bio-geographic regions. The populated part of the Liupan Mountains r...The Liupan Mountains is located in the southern Ningxia Hui Autonomous Region of China, which forms an important dividing line between landforms and bio-geographic regions. The populated part of the Liupan Mountains region has suffered tremendous ecological damages over time due to population pressure, excessive demand and inappropriate use of agricultural land resources. In this paper, datasets of land use between 1990 and 2000 were obtained from Landsat TM imagery, and then spatial models were used to characterize landscape conditions. Also, the relationship between the population density and land use/cover change (LUCC) was analyzed. Results indicate that cropland, forestland, and urban areas have increased by 44,186ha, 9001ha and 1550ha, respectively while the grassland area has appreciably decreased by 54,025ha in the study period. The decrease in grassland was most notable. Of the grassland lost, 49.4% was converted into cropland. The largest annual land conversion rate in the study area was less than 2%. These changes are attributed to industrial and agricultural development and population growth. To improve the eco-economic conditions in the study region, population control, urbanization and development of an ecological friendly agriculture were suggested.展开更多
The basic mathematic models,such as the statistic model,the time-serial model,the spatial dynamic model etc.,and some typical analysis methods based on 3DCM are proposed and discussed.A few typical spatial decision ma...The basic mathematic models,such as the statistic model,the time-serial model,the spatial dynamic model etc.,and some typical analysis methods based on 3DCM are proposed and discussed.A few typical spatial decision making methods integrating the spatial analysis and the basic mathematical models are also introduced,e.g.visual impact assessment,dispersion of noise immissions,base station plan for wireless communication.In addition,a new idea of expectation of further applications and add-in-value service of 3DCM is promoted.As an example,the sunshine analysis is studied and some helpful conclusions are drawn.展开更多
现有相关工程与技术多关注建筑构件几何信息的映射,构件语义信息挖掘侧重于工程项目需要的单一信息,导致BIM(Building Information Modeling)数据在语义交互和智能分析中的潜力未被充分挖掘。该文提出一种基于知识图谱的BIM建筑构件语...现有相关工程与技术多关注建筑构件几何信息的映射,构件语义信息挖掘侧重于工程项目需要的单一信息,导致BIM(Building Information Modeling)数据在语义交互和智能分析中的潜力未被充分挖掘。该文提出一种基于知识图谱的BIM建筑构件语义信息提取方法,首先通过设计语义映射规则将IFC(Industry Foundation Classes)模型的实体、属性、物理信息转化为知识图谱,形成可用于语义分析的结构化语义网络,然后采用TransE嵌入模型对构建的BIM知识图谱进行嵌入学习,通过向量化表示增强信息提取能力。以某三层综合建筑楼为研究对象,提取并构建了包含996个BIM语义节点和2173条关系的知识图谱,进一步采用TransE模型进行语义信息嵌入,对提取结果进行验证。实验结果表明:知识图谱能有效提取BIM建筑构件语义信息,选取最优参数后进行TransE模型嵌入学习,实体语义成功率为97.27%,该方法能够精准捕捉建筑构件各类语义信息的关键内容,减少信息遗漏和提取错误,为BIM模型信息分析和决策提供了新思路。展开更多
界址点与界址线的提取在土地确权工作中是一项非常重要的工作。传统的界址点与界址线提取方法需要经过人工判断,根据判断后的结果将界址点与界址线相对应的属性信息手动输入到属性表中。而过多的人工干预使工作效率降低,同时也容易引发...界址点与界址线的提取在土地确权工作中是一项非常重要的工作。传统的界址点与界址线提取方法需要经过人工判断,根据判断后的结果将界址点与界址线相对应的属性信息手动输入到属性表中。而过多的人工干预使工作效率降低,同时也容易引发各类错误。针对上述问题提出了一种界址点与界址线提取的优化算法。该算法的思路为:(1)利用ArcGIS中的建模工具建立一套针对全部宗地的数据预处理模型,通过该模型生成相应的界址点与界址线的模板文件;(2)利用地理空间数据抽象库(geospatial data abstraction library,GDAL)读取宗地与模板文件的若干属性字段信息;(3)生成并自动读取宗地界址点与界址线文件。实验结果表明,该方法能够有效避免人工干预,提高了界址点与界址线的提取效率。在实际生产尤其是在大批量界址点和界址线的提取工作中,该方法具有相当的优势,为中国同类土地监测与管理工作提供了原创性技术支持。展开更多
三维城市模型是智慧城市与数字孪生城市的核心内容。IFC(industry foundation classes)标准构建的建筑信息模型(building information model,BIM)为城市空间规划和建筑设计提供了丰富而精细的城市三维模型数据。由于标准的差异,BIM无法...三维城市模型是智慧城市与数字孪生城市的核心内容。IFC(industry foundation classes)标准构建的建筑信息模型(building information model,BIM)为城市空间规划和建筑设计提供了丰富而精细的城市三维模型数据。由于标准的差异,BIM无法与三维地理信息系统无缝集成和交互应用,极大地限制了IFC标准BIM的应用优势。本文挖掘IFC、CityGML(city geography markup language)标准三维建筑模型的语义表达差异,提出BIM源模型向不同细节层次(level of detail,LOD)层级模型的几何转换方法;构建一套完整的转换流程,实现IFC标准BIM向CityGML标准三维建筑模型的几何转换;设计并开发转换原型系统。研究表明:本方法及实现流程能够有效实现IFC-BIM模型向CityGML-BIM模型的不同LOD层级的转换,且对多数单体模型相当实用;转换后模型未见错误与偏差,视觉效果良好;具备灵活性,可根据不同模型的转换进行调整。研究成果为BIM与地理信息系统的集成应用提供了新思路,为智慧城市与数字孪生开辟了一种新的数据来源通道。展开更多
文摘视觉同步定位与建图(simultaneous localization and mapping,SLAM)是实现移动机器人自主定位并构建环境地图的关键环节。SLAM技术虽能精确重建环境几何结构,却难以为机器人提供执行复杂任务所需的语义理解能力;建筑信息模型(building information model,BIM)包含丰富的建筑信息,但与机器人操作系统(robot operating system,ROS)之间存在显著的数据格式和表达方式差异,且现有研究多采用人工方式进行转换,效率低下难以规模化应用,且室内环境并非静态不变,从而会影响机器人的导航决策。因此,提出一种集成BIM数据的ROS室内语义地图构建与动态更新方法。通过研发工业基础类(industry foundation classes,IFC)到统一机器人描述格式(unified robot description format,URDF)自动转换器,实现从BIM到机器人仿真环境的自动化建模;融合YOLOv8与随机采样一致性(random sample consensus,RANSAC)算法,建立视觉驱动的语义地图动态更新机制。结果表明,静态建筑元素还原准确率达98%以上,动态物体识别精度达0.9以上,显著提升了语义地图的自动化程度、知识丰富度及环境适应性。
基金Under the auspices of the National Key Science and Technology Support Program of China (No. 2006BCA01A07-2)National Natural Science Foundation of China (No. 40671153)+1 种基金Hunan Land Resource Bureau Program (No. 2007-15)Hunan Educa-tion Bureau Program (No. 08C348)
文摘The Liupan Mountains is located in the southern Ningxia Hui Autonomous Region of China, which forms an important dividing line between landforms and bio-geographic regions. The populated part of the Liupan Mountains region has suffered tremendous ecological damages over time due to population pressure, excessive demand and inappropriate use of agricultural land resources. In this paper, datasets of land use between 1990 and 2000 were obtained from Landsat TM imagery, and then spatial models were used to characterize landscape conditions. Also, the relationship between the population density and land use/cover change (LUCC) was analyzed. Results indicate that cropland, forestland, and urban areas have increased by 44,186ha, 9001ha and 1550ha, respectively while the grassland area has appreciably decreased by 54,025ha in the study period. The decrease in grassland was most notable. Of the grassland lost, 49.4% was converted into cropland. The largest annual land conversion rate in the study area was less than 2%. These changes are attributed to industrial and agricultural development and population growth. To improve the eco-economic conditions in the study region, population control, urbanization and development of an ecological friendly agriculture were suggested.
基金Funded by the National Seientific Foundation of China(No.40001017)the Fok Ying Tung Education Foundation(No.71017)the LIESM ARS Foun-dation(W KL(02)0301)and the Chinese PostdoctoraI Foundation(No.2003033454).
文摘The basic mathematic models,such as the statistic model,the time-serial model,the spatial dynamic model etc.,and some typical analysis methods based on 3DCM are proposed and discussed.A few typical spatial decision making methods integrating the spatial analysis and the basic mathematical models are also introduced,e.g.visual impact assessment,dispersion of noise immissions,base station plan for wireless communication.In addition,a new idea of expectation of further applications and add-in-value service of 3DCM is promoted.As an example,the sunshine analysis is studied and some helpful conclusions are drawn.
文摘现有相关工程与技术多关注建筑构件几何信息的映射,构件语义信息挖掘侧重于工程项目需要的单一信息,导致BIM(Building Information Modeling)数据在语义交互和智能分析中的潜力未被充分挖掘。该文提出一种基于知识图谱的BIM建筑构件语义信息提取方法,首先通过设计语义映射规则将IFC(Industry Foundation Classes)模型的实体、属性、物理信息转化为知识图谱,形成可用于语义分析的结构化语义网络,然后采用TransE嵌入模型对构建的BIM知识图谱进行嵌入学习,通过向量化表示增强信息提取能力。以某三层综合建筑楼为研究对象,提取并构建了包含996个BIM语义节点和2173条关系的知识图谱,进一步采用TransE模型进行语义信息嵌入,对提取结果进行验证。实验结果表明:知识图谱能有效提取BIM建筑构件语义信息,选取最优参数后进行TransE模型嵌入学习,实体语义成功率为97.27%,该方法能够精准捕捉建筑构件各类语义信息的关键内容,减少信息遗漏和提取错误,为BIM模型信息分析和决策提供了新思路。
文摘界址点与界址线的提取在土地确权工作中是一项非常重要的工作。传统的界址点与界址线提取方法需要经过人工判断,根据判断后的结果将界址点与界址线相对应的属性信息手动输入到属性表中。而过多的人工干预使工作效率降低,同时也容易引发各类错误。针对上述问题提出了一种界址点与界址线提取的优化算法。该算法的思路为:(1)利用ArcGIS中的建模工具建立一套针对全部宗地的数据预处理模型,通过该模型生成相应的界址点与界址线的模板文件;(2)利用地理空间数据抽象库(geospatial data abstraction library,GDAL)读取宗地与模板文件的若干属性字段信息;(3)生成并自动读取宗地界址点与界址线文件。实验结果表明,该方法能够有效避免人工干预,提高了界址点与界址线的提取效率。在实际生产尤其是在大批量界址点和界址线的提取工作中,该方法具有相当的优势,为中国同类土地监测与管理工作提供了原创性技术支持。
文摘三维城市模型是智慧城市与数字孪生城市的核心内容。IFC(industry foundation classes)标准构建的建筑信息模型(building information model,BIM)为城市空间规划和建筑设计提供了丰富而精细的城市三维模型数据。由于标准的差异,BIM无法与三维地理信息系统无缝集成和交互应用,极大地限制了IFC标准BIM的应用优势。本文挖掘IFC、CityGML(city geography markup language)标准三维建筑模型的语义表达差异,提出BIM源模型向不同细节层次(level of detail,LOD)层级模型的几何转换方法;构建一套完整的转换流程,实现IFC标准BIM向CityGML标准三维建筑模型的几何转换;设计并开发转换原型系统。研究表明:本方法及实现流程能够有效实现IFC-BIM模型向CityGML-BIM模型的不同LOD层级的转换,且对多数单体模型相当实用;转换后模型未见错误与偏差,视觉效果良好;具备灵活性,可根据不同模型的转换进行调整。研究成果为BIM与地理信息系统的集成应用提供了新思路,为智慧城市与数字孪生开辟了一种新的数据来源通道。
