共查询到20条相似文献，搜索用时 292 毫秒
1.
结合多分类器的遥感数据专题分类方法研究 总被引：19，自引：1，他引：19
采用标准的多分类器结合方法进行遥感图像的分类研究。首先介绍了标准的多分类器结合的算法,然后以Landsat-TM多光谱遥感数据的土地覆被分类为例,分别给出了抽象级上相同训练特征的多分类器结合、抽象级上不同训练特征的多分类器结合和测量级上的多分类器结合进行土地覆被分类的方法,并进行了实例研究。参与分类器结合的单个分类器包括最大似然分类器,最小距离分类器,马氏距离分类器,K-NN分类器,多层感知器神经网络分类器。分类器的分类精度用总体精度、用户精度、生产者精度、kappa系数和条件kappa系数评价。结果表明,每一种多分类器结合的分类方法都能够比较显著地提高总体分类精度。文章最后对不同多分类器结合方式的优缺点进行了分析。 相似文献
2.
Mhosisi Masocha Andrew K. Skidmore 《International Journal of Applied Earth Observation and Geoinformation》2011
Mapping the cover of invasive species using remotely sensed data alone is challenging, because many invaders occur as mid-level canopy species or as subtle understorey species and therefore contribute little to the spectral signatures captured by passive remote sensing devices. In this study, two common non-parametric classifiers namely, the neural network and support vector machine were used to map four cover classes of the invasive shrub Lantana camara in a protected game reserve and the adjacent area under communal land management in Zimbabwe. These classifiers were each combined with a geographic information system (GIS) expert system, in order to test whether the new hybrid classifiers yielded significantly more accurate invasive species cover maps than the single classifiers. The neural network, when used on its own, mapped the cover of L. camara with an overall accuracy of 71% and a Kappa index of agreement of 0.61. When the neural network was combined with an expert system, the overall accuracy and Kappa index of agreement significantly increased to 83% and 0.77, respectively. Similarly, the support vector machine achieved an overall accuracy of 64% with a Kappa index of agreement of 0.52, whereas the hybrid support vector machine and expert system classifier achieved a significantly higher overall accuracy of 76% and a Kappa index of agreement of 0.67. These results suggest that integrating conventional image classifiers with an expert system increases the accuracy of invasive species mapping. 相似文献
3.
基于支持向量机的无限集成学习方法(SVM-based IEL)是机器学习领域新兴起的一种集成学习方法。本文将SVM-based IEL引入遥感图像的分类领域,并同时将SVM、Bagging、AdaBoost和SVM-based IEL等方法应用于遥感图像分类。实验表明:Bagging方法可以提高遥感图像的分类精度,而AdaBoost却降低了遥感图像的分类精度;同时,与SVM、有限集成的学习方法相比,SVM-based IEL方法具有可以显著地提高遥感图像的分类精度、分类效率的优势。 相似文献
4.
Mbulisi Sibanda Timothy Dube Victor M. Bangamwabo Onisimo Mutanga Cletah Shoko Webster Gumindoga 《国际地球制图》2016,31(9):1006-1018
The objective of this study was to understand the factors that explain the spatial distribution of elephant poaching activities in the areas of the mid-Zambezi Valley, Zimbabwe using geographic information system (GIS) and remotely sensed data integrated with spatial logistic regression. The results showed that significant (α = 0.05) elephant poaching hot spots are located closer to wildlife protected areas. Results further demonstrated that resource availability (water and forage) are the main factors explaining elephant poaching activities in the mid-Zambezi Valley. For example, the majority of poaching activities were found to occur in areas with high vegetation fractional cover (high forage) and close to waterholes. The results also showed that poaching incidences were more prevalent during the dry season. The findings of this study highlight the significance of integrating GIS, remotely sensed data and spatial logistic regression tools for understanding and monitoring elephant poaching activities. This information is critical if poaching activities are to be minimized and it is also important for planning, monitoring and mitigation of poaching activities in similar protected areas across the sub-Saharan Africa. 相似文献
5.
基于自适应共振模型的遥感影像分类方法研究 总被引：9，自引：1，他引：9
人工神经网络（ANN）是人视觉和服的基本功能的抽象、简化和模拟。在对遥感影像的综合解释应用中，与传统的统计方法和符号逻辑方法相比较，ANN更接近人对影像的视觉解译分析过程。自适应共振理论（ART）是一种自组织产生认知编码的神经网络理论，其自组织、反馈式增量学习机能，能兼顾适应性和稳定性，克服了一般神经网络学习速度慢、网络结构难以确定、局部最小陷阱等缺陷。以FUZZY－ART和ARTMAP为基础，提出基于ART遥感影像非监督和监督分类的一般模型，并以实际上土覆盖分类和城市结构信息提取为应用实例，通过与传统统计方法和一般ANN分类器相比较，ART具有正确率更同、学习速度快、自适应性等优点，是复杂数据分类和信息提取的有效工具。 相似文献
6.
Estimating tropical biomass is critical for establishment of conservation inventories and landscape monitoring. However, monitoring biomass in a complex and dynamic environment using traditional methods is challenging. Recently, biomass estimates based on remotely sensed data and ecological variables have shown great potential. The present study explored the utility of remotely sensed data and topo-edaphic factors to improve biomass estimation in the Eastern Arc Mountains of Tanzania. Twenty-nine vegetation indices were calculated from RapidEye data, while topo-edaphic factors were taken from field measurements. Results showed that using topo-edaphic variables or vegetation indices, biomass could be predicted with an R2 of 0.4. A combination of topo-edaphic variables and vegetation indices improved the prediction accuracy to an R2 of 0.6. Results further showed a decrease in biomass estimates from 1162 ton ha?1 in 1980 to 285.38 ton ha?1 in 2012. This study demonstrates the value of combining remotely sensed data with topo-edaphic variables in biomass estimation. 相似文献
7.
《International Journal of Digital Earth》2013,6(4):330-346
Abstract In recent years, the rough set (RS) method has been in common use for remote-sensing classification, which provides one of the techniques of information extraction for Digital Earth. The discretization of remotely sensed data is an important data preprocessing approach in classical RS-based remote-sensing classification. Appropriate discretization methods can improve the adaptability of the classification rules and increase the accuracy of the remote-sensing classification. To assess the performance of discretization methods this article adopts three indicators, which are the compression capability indicator (CCI), consistency indicator (CI), and number of the cut points (NCP). An appropriate discretization method for the RS-based classification of a given remotely sensed image can be found by comparing the values of the three indicators and the classification accuracies of the discretized remotely sensed images obtained with the different discretization methods. To investigate the effectiveness of our method, this article applies three discretization methods of the Entropy/MDL, Naive, and SemiNaive to a TM image and three indicators for these discretization methods are then calculated. After comparing the three indicators and the classification accuracies of the discretized remotely sensed images, it has been found that the SemiNaive method significantly reduces large quantities of data and also keeps satisfactory classification accuracy. 相似文献
8.
Land cover and land use classification performance of machine learning algorithms in a boreal landscape using Sentinel-2 data 总被引：1，自引：0，他引：1
Abdulhakim Mohamed Abdi 《地理信息系统科学与遥感》2020,57(1):1-20
ABSTRACTIn recent years, the data science and remote sensing communities have started to align due to user-friendly programming tools, access to high-end consumer computing power, and the availability of free satellite data. In particular, publicly available data from the European Space Agency’s Sentinel missions have been used in various remote sensing applications. However, there is a lack of studies that utilize these data to assess the performance of machine learning algorithms in complex boreal landscapes. In this article, I compare the classification performance of four non-parametric algorithms: support vector machines (SVM), random forests (RF), extreme gradient boosting (Xgboost), and deep learning (DL). The study area chosen is a complex mixed-use landscape in south-central Sweden with eight land-cover and land-use (LCLU) classes. The satellite imagery used for the classification were multi-temporal scenes from Sentinel-2 covering spring, summer, autumn and winter conditions. Using stratified random sampling, each LCLU class was allocated 1477 samples, which were divided into training (70%) and evaluation (30%) subsets. Accuracy was assessed through metrics derived from an error matrix, but primarily overall accuracy was used in allocating algorithm hierarchy. A two-proportion Z-test was used to compare the proportions of correctly classified pixels of the algorithms and a McNemar’s chi-square test was used to compare class-wise predictions. The results show that the highest overall accuracy was produced by support vector machines (0.758 ± 0.017), closely followed by extreme gradient boosting (0.751 ± 0.017), random forests (0.739 ± 0.018), and finally deep learning (0.733 ± 0.0023). The Z-test comparison of classifiers showed that a third of algorithm pairings were statistically different. On a class-wise basis, McNemar’s test results showed that 62% of class-wise predictions were significant from one another at the 5% level or less. Variable importance metrics show that nearly half of the top twenty Sentinel-2 bands belonged to the red edge (25%) and shortwave infrared (23%) portions of the electromagnetic spectrum, and were dominated by scenes from spring (38%) and summer (40%). The results are discussed within the scope of recent studies involving machine learning and Sentinel-2 data and key knowledge gaps identified. The article concludes with recommendations for future research. 相似文献
9.
提出了一种基于误差分析的组合分类器,通过结合两种监督分类方法,提出的算法分别估计了两种监督分类方法在计算过程中的误差,给出了规则输出的置信区间,再根据置信区间的大小对两种分类方法的输出结果进行加权平均,从而得到更精确的规则输出.利用该方法对遥感图像进行分类实验,在不同训练样本分布与不同训练样本数量的情况下,比较新的组合分类器与单一分类器的精度.结果表明新的组合分类器能够取得比单一的分类器更高的分类精度.结果还显示出,两个分类器的独立性越强,组合分类器的效果越好.另外一个实验比较了新的组合分类器与和式规则组合分类器的分类精度,结果仍显示出了新方法的优越性. 相似文献
10.
This study investigated the combined use of multispectral/hyperspectral imagery and LiDAR data for habitat mapping across parts of south Cumbria, North West England. The methodology adopted in this study integrated spectral information contained in pansharp QuickBird multispectral/AISA Eagle hyperspectral imagery and LiDAR-derived measures with object-based machine learning classifiers and ensemble analysis techniques. Using the LiDAR point cloud data, elevation models (such as the Digital Surface Model and Digital Terrain Model raster) and intensity features were extracted directly. The LiDAR-derived measures exploited in this study included Canopy Height Model, intensity and topographic information (i.e. mean, maximum and standard deviation). These three LiDAR measures were combined with spectral information contained in the pansharp QuickBird and Eagle MNF transformed imagery for image classification experiments. A fusion of pansharp QuickBird multispectral and Eagle MNF hyperspectral imagery with all LiDAR-derived measures generated the best classification accuracies, 89.8 and 92.6% respectively. These results were generated with the Support Vector Machine and Random Forest machine learning algorithms respectively. The ensemble analysis of all three learning machine classifiers for the pansharp QuickBird and Eagle MNF fused data outputs did not significantly increase the overall classification accuracy. Results of the study demonstrate the potential of combining either very high spatial resolution multispectral or hyperspectral imagery with LiDAR data for habitat mapping. 相似文献
11.
《International Journal of Digital Earth》2013,6(2):115-142
Abstract On site observation is the most common way of monitoring cultural heritage sites and monuments in Cyprus. However, this procedure that includes data collection, periodical observations, and multivariate risk assessment analysis is difficult to accomplish with the traditional practices and methods since it is time consuming and expensive. Furthermore, many archaeological sites and monuments are located at inaccessible areas, far away from the main road network and urban areas. Satellite remote sensing and Geographical Information Systems (GIS) can successfully confront this problem by providing the scientists with integrated monitoring of the study areas and the unique advantage to store and manipulate a large amount of spatial and attribute data simultaneously. Actually the monitoring and identification of several natural and anthropogenic hazards in the vicinity of the cultural heritage sites in Cyprus, seems to be one of the main priorities of its governmental and municipal authorities. This study aims to integrate both satellite remote sensing techniques and GIS in a multidisciplinary approach, for monitoring anthropogenic and natural hazards with the use of archived and up-to-date multitemporal remotely sensed images in the study area, namely in areas nearby cultural heritage sites and monuments in Cyprus. In this study anthropogenic hazards include urbanisation and extended land use changes in the surroundings of archaeological sites and natural hazards concern seismicity and sea erosion. 相似文献
12.
The reliability of habitat maps that have been generated using Geographic Information Systems (GIS) and image processing of remotely sensed data can be overestimated. Habitat suitability and spatially explicit population viability models are often based on these products without explicit knowledge of the effects of these mapping errors on model results. While research has considered errors in population modeling assumptions, there is no standardized method for measuring the effects of inaccuracies resulting from errors in landscape classification. Using landscape‐scale maps of existing vegetation developed for the USDA Forest Service in southern California from Landsat Thematic Mapper satellite data and GIS modeling, we performed a sensitivity analysis to estimate how mapping errors in vegetation type, forest canopy cover, and tree crown size might affect delineation of suitable habitat for the California spotted owl (Strix occidentalis occidentalis). The resulting simulated uncertainty maps showed an increase in the estimated area of suitable habitat types. Further analysis measuring the fragmentation of the additional patches showed that they were too small to be useful as habitat areas. 相似文献
13.
Abstract This paper investigates the combination of metric aerial photography and near‐infrared (NIR) videography data to improve the design of field‐survey sampling frameworks. Spatial data collection can contribute up to 80% of the cost of deploying a Geographic Information System (GIS) based Decision Support System (DSS). The use of remotely sensed information, field survey using differential Global Positioning System (dGPS) and geostatistical interpolation methods maximises data quality for a given rate of sampling. Medium‐format colour aerial photography and NIR videography were orthorectified to the national map base and mosaiced using ERDAS Imagine. The green and red layers of the aerial photography were combined with the NIR videography to form a false‐colour composite image. Two sampling strategies were tested. The first stratified sampling on a per field basis, creating four points per hectare, randomly located within each field. The second strategy used the remotely sensed information to identify within‐field variability classes for each field, using red‐green difference or normalised difference vegetation index (NDVI) models. These variability classes were used as a sub‐stratification framework with each class sampled at the same rate of 4 per hectare. For both strategies the sample points were generated within ESRI ArcView and were located in the field using dGPS. Maps of stone content were created using geostatistical methods and validated against samples collected on a 100 metre grid. It was concluded that combining the two image sources to create a within‐field stratification framework improved the precision of the results obtained from field‐survey. 相似文献
14.
Forrest R. Stevens Andrea E. Gaughan Jeremiah J. Nieves Adam King Alessandro Sorichetta Catherine Linard 《International Journal of Digital Earth》2020,13(1):78-100
ABSTRACTMapping built land cover at unprecedented detail has been facilitated by increasing availability of global high-resolution imagery and image processing methods. These advances in urban feature extraction and built-area detection can refine the mapping of human population densities, especially in lower income countries where rapid urbanization and changing population is accompanied by frequently out-of-date or inaccurate census data. However, in these contexts it is unclear how best to use built-area data to disaggregate areal, count-based census data. Here we tested two methods using remotely sensed, built-area land cover data to disaggregate population data. These included simple, areal weighting and more complex statistical models with other ancillary information. Outcomes were assessed across eleven countries, representing different world regions varying in population densities, types of built infrastructure, and environmental characteristics. We found that for seven of 11 countries a Random Forest-based, machine learning approach outperforms simple, binary dasymetric disaggregation into remotely-sensed built areas. For these more complex models there was little evidence to support using any single built land cover input over the rest, and in most cases using more than one built-area data product resulted in higher predictive capacity. We discuss these results and implications for future population modeling approaches. 相似文献
15.
This study tested the use of machine learning techniques for the estimation of above-ground biomass (AGB) of Sonneratia caseolaris in a coastal area of Hai Phong city, Vietnam. We employed a GIS database and multi-layer perceptron neural networks (MLPNN) to build and verify an AGB model, drawing upon data from a survey of 1508 mangrove trees in 18 sampling plots and ALOS-2 PALSAR imagery. We assessed the model’s performance using root-mean-square error, mean absolute error, coefficient of determination (R2), and leave-one-out cross-validation. We also compared the model’s usability with four machine learning techniques: support vector regression, radial basis function neural networks, Gaussian process, and random forest. The MLPNN model performed well and outperformed the machine learning techniques. The MLPNN model-estimated AGB ranged between 2.78 and 298.95 Mg ha?1 (average = 55.8 Mg ha?1); below-ground biomass ranged between 4.06 and 436.47 Mg ha?1 (average = 81.47 Mg ha?1), and total carbon stock ranged between 3.22 and 345.65 Mg C ha?1 (average = 64.52 Mg C ha?1). We conclude that ALOS-2 PALSAR data can be accurately used with MLPNN models for estimating mangrove forest biomass in tropical areas. 相似文献
16.
尾矿库是矿山企业选矿的必要设施,同时对周围环境也是一个重大的危险源。为研究流域范围内尾矿库的溃坝路径以及对矿区地表造成的风险,以赤城县为例,利用GF-1高分辨率遥感影像,基于遥感RS（Remote Sensing）及地理信息系统GIS（Geographic Information System）手段对尾矿库的流域风险进行了监测分析。首先,通过分析尾矿库在遥感影像上的纹理、色调、形状以及大小等特征,制作了用于目标检测的尾矿库样本集,然后,在原始SSD目标检测网络基础上添加了反卷积模块和连接模块构建多尺度融合目标检测算法MSF_SSD,在目标检测结果基础上使用PSPnet算法实现尾矿库结构分割,得到了尾矿库内部结构—坝体以及库区,运用RS与GIS技术对尾矿库的上游汇水面以及事故可能径流进行提取,进而基于Arc Hydro模型模拟尾矿库的溃坝路径。最后,通过构建溃坝路径的缓冲区,得到尾矿库发生溃坝所造成的地物影响范围及面积。研究结果表明：赤城县尾矿库的溃坝路径总体是从西向东,从北向南,受溃坝影响的地物总面积达到480 km2。其中,林地176.52 km2,耕地175.52 km2,城市建设用地43.74 km2,农村建设用地2.47 km2,水体17.72 km2,草地和牧场分别为3.60 km2、1.22 km2。研究成果可用于分析尾矿库溃坝造成的地物损失以及影响范围面积等信息,提升尾矿库的风险管理水平及应急响应能力,为有关部门制定决策提供理论依据。 相似文献
17.
《The Cartographic journal》2013,50(1):21-26
AbstractThe spatially discontinuous choropleth map is a poor representation of the underlying continuous distribution of population density. A possible alternative is to derive dasymetric maps at a fine spatial resolution by making use of satellite imagery in a geographical information system. However, there are cartographic problems when these maps are displayed and further processing is needed in order to obtain approximations to a continuous density surface. Isarithmic maps of these density surfaces retain a high degree of spatial accuracy while providing pleasing and highly adaptable presentations.The methods used to generate dasymetric and isarithmic maps are readily implemented in most raster based geographical information systems. For example, the classification of remotely sensed imagery, the subsequent processing and integration of data, and most of the cartographic display, were all undertaken in this work using the low cost IDRISI GIS that operates on standard IBM PC compatible hardware. 相似文献
18.
Remote sensing has recently been used to map forage quality for rangeland management. However, the validity of the remotely sensed forage quality can best be assessed when it connects well with the animal unit. In this study we used the new WorldView-2 multispectral imagery to estimate and map forage quality (nitrogen concentration) as a step to explain GPS based cattle distribution in a rangeland of Southeastern Zimbabwe. Nitrogen concentration was successfully estimated and mapped (Rcv2 = 0.66, relative error = 0.13%) using partial least squares regression (PLSR). The integration of GPS based cattle distribution patterns with forage quality in a GIS showed that cattle locations significantly clustered in areas of high forage quality. The results of this study suggest that new multispectral data with unique band settings such as WorldView-2 improves the estimation and mapping of forage quality in rangelands at landscape level. In addition, our results indicate that remotely sensed forage quality can be used to explain herbivore distribution, particularly cattle grazing patterns in rangelands. 相似文献
19.
20.
The aim of this study is to compare the changes that occurred in the main urban land-cover classes of Ulaanbaatar city, Mongolia, during a centralized economy with those that occurred during a market economy and to describe the socio-economic reasons for the changes. For this purpose, multi-temporal remote sensing and geographical information system (GIS) data sets, as well as census data, are used. To extract the reliable urban land-cover information from the selected remotely sensed data sets, a refined parametric classification algorithm that uses spatial thresholds defined from local and contextual knowledge is constructed. Before applying the classification decision rule, some image fusion techniques are applied to the selected remotely sensed data sets to define the most efficient fusion method for training sample selection and for defining local and contextual knowledge. Overall, the study indicates that during the centralized economy significant changes occurred in a ger area of the city, whereas during the market economy the changes occurred in all areas. 相似文献