共查询到20条相似文献，搜索用时 203 毫秒
1.
基于端到端的检测框架,使用多空洞率卷积核组作为特征提取模块,并在不同特征提取层间设置了密集连接,来加强不同尺度特征图内的信息复杂度;以多尺度特征图融合为基础,构建了4个输出层的特征图上采样金字塔,最后通过数据增强提高了训练集内目标的表达能力。测试结果表明,本文方法在测试数据集上能够达到较高的检测精度,体现了良好的实时检测能力,并且对不同背景下多角度的房屋目标具有很好的泛化性能。该方法在城市违章建筑监管与智慧城市建设等领域具有较高的实用价值。 相似文献
2.
3.
高光谱遥感影像(hyperspectral imagery, HSI)通常包含几十至数百个连续波段,具有图谱合一、光谱连续的特点,能够实现地物的精细分类,被广泛应用农业、林业、城市以及海洋等领域。HSI特征提取是高光谱应用的前提,也是遥感领域的研究热点和前沿课题之一。近年来,奇异谱分析(singular spectrum analysis, SSA)被应用于HSI领域,在光谱特征和空间特征提取方面取得了较好效果,逐渐成为特征提取的一种有效方法。本文首先分析了HSI特征提取的研究进展和存在的问题;其次对SSA方法进行了系统的梳理,分别介绍了光谱域1D-SSA、空间域2D-SSA和光谱-空间组合域SSA 3类方法的作用、效果及优缺点,并在两个公开的HSI数据集和一个高分五号HSI数据上进行了分类效果验证;最后,对SSA特征提取进行了总结,并讨论了未来的研究方向。 相似文献
4.
5.
基于车载激光扫描数据的目标分类方法 总被引：4，自引：2，他引：2
车裁扫描系统可以获取建筑物、道路、隧道等城市物体的表面信息,非常适用于城市物体三维空间信息的快速有效获取。本文从激光扫描数据中进行建筑物特征提取,重点是车载激光扫描数据分类;设计了一种新的车载激光扫描数据分类和建筑物特征提取方法,此方法充分利用激光扫描数据的空间分布特征和城市环境中各种物体自身的几何特征,将所有数据点投影到水平格网中,对每个格网单元,计算其中所有数据点投影前高度值的最大值,根据此最大值来判定该格网单元中数据点的类别,从分类后的数据中提取建筑物特征。应用所设计的数据处理方法处理车载式三维信息采集系统采集的数据,分析得到了结果。 相似文献
6.
高光谱遥感影像丰富的光谱信息有利于深入挖掘目标的理化特性,精细识别不同目标间的细微差异。为了提高影像分类识别的精度与速度需要对光谱信息进行特征提取。基于核函数的判别分析能够在数据中提取非线性特征,本文将其应用到高光谱影像分类的特征提取中,并进行了最小距离分类实验取得理想结果。 相似文献
7.
合成孔径雷达(SAR)图像含有丰富的纹理信息,特别是进行城市地物分类时,纹理特征对于图像的解译具有重要的意义。本文对基于灰度共生矩阵和Gabor变换两种纹理特征提取方法进行了研究,将灰度和不同纹理特征组合应用于SAR图像城市地物分类,并以ALOS PALSAR影像为数据源进行了实验。通过对不同分类结果进行定性和定量分析,结果表明,引入纹理特征后的SAR图像分类结果要优于无纹理信息参与的分类结果,基于不同纹理特征组合的SAR图像分类结果要优于基于单一纹理特征的分类结果。 相似文献
8.
合成孔径雷达（ SAR）图像含有丰富的纹理信息,特别是进行城市地物分类时,纹理特征对于图像的解译具有重要的意义。本文对基于灰度共生矩阵和Gabor变换两种纹理特征提取方法进行了研究,将灰度和不同纹理特征组合应用于SAR图像城市地物分类,并以ALOS PALSAR影像为数据源进行了实验。通过对不同分类结果进行定性和定量分析,结果表明,引入纹理特征后的SAR图像分类结果要优于无纹理信息参与的分类结果,基于不同纹理特征组合的SAR图像分类结果要优于基于单一纹理特征的分类结果。 相似文献
9.
特征提取对建筑物精细建模的品质和精度起着重要作用。为清晰准确地提取建筑物的特征信息,本文针对采用传统的法矢估计方法受噪声影响大、存在误判的问题,提出了一种基于移动最小二乘法矢估计的建筑物点云特征提取方法。该方法首先采用移动最小二乘法进行法矢估计,然后将K邻域法矢夹角的均值作为点的显著性指标进行特征点判别,最后对提取出的特征点集进行下采样,进一步消除冗余信息。试验结果表明,采用移动最小二乘法进行点云法矢估计,其结果更加准确和稳健,从而有效提升了建筑物点云特征提取的精确性和可靠性,对特征点集的下采样能够删除大量冗余特征点,使提取的特征线更加简洁、清晰、完整。 相似文献
10.
传统城市特征识别采用空间和统计方法提取分析指标,特征评价指标主观性较大。街景影像包含城市视觉信息,可以进行城市特征识别。以中国青岛市为例,构建面向街景的多尺度语义分割模型MS-DeepLabV3+。在编码区增加全特征提取通道聚合多尺度特征;在解码区增加多尺度特征提取通道,有效捕捉低层次特征;引入注意力机制模块和通道注意力,聚焦关键特征,提高街景语义分割的准确性,模型平均交并比、精确率和召回率分别提高了3.47%、2.37%和3.96%。在地块尺度上,从6个维度建立了城市多维特征向量,即环境维度、设施便利维度、经济富裕度、交通维度、城市安全维度和城市综合度,结合兴趣点数据和居住用地数据,以表征青岛市各城区的城市特征。使用Grad-CAM方法对语义分割模型进行可解释分析,采用特征归因SHAP方法挖掘了城市多维特征的内在驱动因素。结果发现,不同城区具有不同的特征向量,不同城区的特征向量具有在特定维度上的优势。研究结果有助于优化城市空间中多维度特征,为城市的规划建设提供参考。 相似文献
11.
The electric vehicle industry is essential for China's economic transformation and industrial upgrading. Identifying the distribution of electric vehicle sales and service shops is of great significance to boost the development of retail geography theory and promote the optimization of industrial space. Regression and spatial econometric methods analyze the influencing factors of electric vehicle sales and service shop distribution. According to the point of interest data of electric vehicle sales and service shops, taking 10 typical cities in China as an example, the spatial distribution of electric vehicle sales and service shops and its influencing factors are empirically explored by spatial econometric methods. The results indicate the following: (1) From high to low, the influence degree of spatial dependence factors are car license restriction policy, the average age of residents, number of schools, built-up area, the amount of investment in real estate, electric vehicle sales, and GDP; (2) On the whole, the spatial distribution of electric vehicle sales and service shops shows the characteristics of a multi-center circle-layered spatial structure that spreads from the inside to the outside; and (3) The spatial influencing factors such as electric vehicle sales, charging stations, permanent resident population, schools, and subway stations in the four first-tier cities is evident. 相似文献
12.
Tomoki Nakaya A. Stewart Fotheringham Kazumasa Hanaoka Graham Clarke Dimitris Ballas Keiji Yano 《Journal of Geographical Systems》2007,9(4):345-369
Although the disaggregation of consumers is crucial in understanding the fragmented markets that are dominant in many developed
countries, it is not always straightforward to carry out such disaggregation within conventional retail modelling frameworks
due to the limitations of data. In particular, consumer grouping based on sampled data is not assured to link with the other
statistics that are vital in estimating sampling biases and missing variables in the sampling survey. To overcome this difficulty,
we propose a useful combination of spatial interaction modelling and microsimulation approaches for the reliable estimation
of retail interactions based on a sample survey of consumer behaviour being linked with other areal statistics. We demonstrate
this approach by building an operational retail interaction model to estimate expenditure flows from households to retail
stores in a local city in Japan, Kusatsu City.
相似文献
13.
Creating a GIS application for retail centers in Jeddah city 总被引：2，自引：0，他引：2
AbdulKader A. Murad 《International Journal of Applied Earth Observation and Geoinformation》2003,4(4):329-338
The aim of this paper is to identify the role of geographical information systems (GIS) in supporting retail planners in monitoring and analyzing retail development and growth. At first, a review about retail planning and its relevant issues is made. Second, GIS is defined together with its applications in retail planning. The third part discusses the created GIS application that is designed for two retail centers in Jeddah city, Saudi Arabia. This application is created to help planners in defining the distribution of retail demand, describing customers profile, analyzing trade areas and modeling retail flows. Two useful models have been produced by this study. One is called market penetration and the other is based on spatial interaction technique. The former is created to examine retail center catchment area, while the latter is used to model the interaction between demand areas and retail centers. 相似文献
14.
传统零售商业网点选址模型主要是基于统计的静态模型。分析了影响零售商业网点选址的人口、经济、市场竞争三个关键要素,结合GIS,提出了一种基于权重设置的零售商业网点选址评价模型,并阐述了模型的应用流程。该模型在深圳市罗湖区公共信息服务平台中进行了典型应用实现,并检验了模型的应用可行性。 相似文献
15.
《International Journal of Digital Earth》2013,6(5):403-420
The use of spatial data has become a key factor in numerous public sector processes. However, recent evidence suggests that the seemingly obvious benefits of spatial data use and sharing often stay beyond reach. This paper examines whether the performance of spatial data in inter-organisational processes depends on the structural characteristics of the process involved. Process structure here refers to the way in which tasks are divided and coordinated between the organisations in a process chain, and to the allocation of the spatial data handling tasks in that process. Spatial data performance describes the level of access, use and sharing of spatial data in the process. The paper presents the results of a comparative case study of inter-organisational public sector processes in the fields of urban planning, flood management, population registration, environmental permit delivery and road safety monitoring. The findings suggest that spatial data performance in inter-organisational processes depends on the interplay between the structural characteristics of the process. Unfragmented processes with embedded activities that are related to spatial data are apparently linked to a high level of spatial data performance, while fragmented processes without centralised coordination, or processes in which the spatial data-related activities are separated from the actual process, seem to relate to lower levels of spatial data performance. 相似文献
16.
Investigating the association between streetscapes and human walking activities using Google Street View and human trajectory data 总被引：1，自引：0，他引：1 
下载免费PDF全文
下载免费PDF全文 Xiaojiang Li Paolo Santi Theodore K. Courtney Santosh K. Verma Carlo Ratti 《Transactions in GIS》2018,22(4):1029-1044
Having an active lifestyle is recognized to positively contribute to public health. Creating more walkable streets and neighborhoods is an important way to promote an active lifestyle for urban residents. It is therefore important to understand how the urban built environment can influence human walking activities. In this study, we investigated the interaction of human walking activities and physical characteristics of streetscapes in Boston. A large number of anonymous pedestrian trajectories collected from a smartphone application were used to estimate human walking activities. Publicly accessible Google Street View images were used to estimate the amount of street greenery and the enclosure of street canyons, both of which were used to indicate the physical characteristics of streetscapes. The Walk Score and population were also added in the statistical analyses to control the influence of nearby urban facilities and population on human walking activities. Statistical analysis results show that both the street greenery and the enclosure of the street canyons are significantly associated with human walking activities. The associations between the streetscape variables and human walking activities vary in different land use types. The results of this study have implications for designing walkable and healthy cities. 相似文献
17.
A spatio-temporal model of housing price trends is developed that focuses on individual housing sales over time. The model
allows for both the spatio-temporal lag effects of previous sales in the vicinity of each housing sale, and for general autocorrelation
effects over time. A key feature of this model is the recognition of the unequal spacing between individual housing sales
over time. Hence the residuals are modeled as a first-order autoregressive process with unequally spaced events. The maximum-likelihood
estimation of this model is developed in detail, and tested in terms of simulations based on selected data. In addition, the
model is applied to a small data set in the Philadelphia area. 相似文献
18.
When planners intervene in urban systems they seek in part to generate a positive response from those parts of a city outside of their control, namely the private market and private individuals. This response is difficult to predict due to the fact that the city is a nonlinear system of organized complexity. Models of cities which seek to explain this response are necessarily complicated and dynamic. Where an analytical solution is not possible we turn to computer simulation and interactive visualization in order to understand their output. Allowing human participation in such simulations provides a sandbox in which to experiment with the dynamic behaviour of an urban model and play a part in its evolution. Two possible options for structuring this participation are: (1) toy retail systems, which allow unconstrained experimentation, and (2) games, which impose rules and involve role‐play and competition. To explore these ideas we construct a toy retail system and a two‐player retail game, both of which are derived from an existing agent‐based retail model. We explore the application of these systems to the metropolitan county of South Yorkshire in the UK. 相似文献
19.
There is considerable interest in understanding the distribution patterns of different types of retail space, over time, and doing so at a national scale. Yet a lack of suitable data, coupled with poor classification schemas, has stymied efforts to create such a national perspective. This research reports on metrics and classification methodologies that have been applied to large scale topographic data, that afford a systematic classification of certain retail spaces potentially at the national coverage. By analysing the form, composition, extent and patterns of buildings within retail spaces, together with their degree of centrality and levels of access, we demonstrate that it is possible to classify different types of retail space. The research illustrates the utility of fine scale topographic data beyond mere mapping. The article compares three methodologies used for classification (Boolean, fuzzy logic and Bayesian modelling) and evaluates them through comparison with known locations of various retail types as a way of assessing the validity of these approaches. The quality of the results are good, though the work highlights the inconsistency in definitions that currently exist – reflecting, as much as anything, the shifting sands of definitions of various retail spaces that ebb and flow according to consumer needs, and the ambitions of urban planners. 相似文献
20.
无人机航拍影像具有分辨率高、回访周期短等特点,利用无人机遥感技术手段对城市范围的建设进行动态监测,可及时、有效地发现涉嫌违法的建设活动。本文结合实际项目需求,研究通过卷积神经网络方法进行违章建筑的自动检测,替代过去靠大量人力检查的模式,目前测试区域无人机影像试验取得了较好的效果,在样本数据不足5000份的情况下,准确率和召回率分别达到了71%和88%。随着样本数据的不断增多,基于该深度学习方法将较大程度上持续提升检测准确率和召回率,能够更精准地发现违法活动,具有较大的实际应用价值及潜力。 相似文献