An evaluation of Radarsat-2 individual and combined image dates for land use/cover mapping

Various land use/cover types exhibit seasonal characteristics which can be captured in remotely sensed imagery. This study examined how different seasons of Radarsat-2 data influence land use/cover classification accuracies for two study sites. Two dates of Radarsat-2 C-band quad-polarised images were obtained for Washington, DC, USA and Wad Madani, Sudan. Spectral signatures were extracted and used with a maximum likelihood decision rule for classification and thematic accuracies were then determined. Both despeckled radar and derived texture measures were examined. Thematic accuracies for the two despeckled image dates were similar with a difference of 3% for Washington and 6% for Sudan. Merging the despeckled images for both seasons increased overall accuracy by 2% for Washington and 9% for Sudan. Further combining the original radar for both seasons with derived texture measures increased overall accuracies by 9% for Washington and 16% for Sudan for final overall accuracy values of 73 and 82%.     

基于光谱和纹理特征的山区高分辨率遥感影像分类

本文在只做阴影补偿而不做地形校正的情况下,使用光谱和纹理特征相结合的方法进行山区高分辨率遥感影像分类。实验取得了78%的分类精度,表明该方法合理可行,具有一定的实用性。     

Landsat mss data for land use and land cover inventory and mapping: A case study of Andhra Pradesh

This study deals with the technique of remote sensing and how far it helps in the rapid study of geographical phenomena especially land use within a very short time and accurate manner. It evaluates how well data from the Landsat - Multispectral Scanner (MSS) could be used to detect, identify and delineate land use features within the Andhra Pradesh State. The main objective was to prepare a small scale land use map from satellite imagery showing the broad distribution of land use patterns to serve as a base for monitoring land use change.     

Spatial and semantic effects of LUCAS samples on fully automated land use/land cover classification in high-resolution Sentinel-2 data

In this study, we test the use of Land Use and Coverage Area frame Survey (LUCAS) in-situ reference data for classifying high-resolution Sentinel-2 imagery at a large scale. We compare several pre-processing schemes (PS) for LUCAS data and propose a new PS for a fully automated classification of satellite imagery on the national level. The image data utilizes a high-dimensional Sentinel-2-based image feature space. Key elements of LUCAS data pre-processing include two positioning approaches and three semantic selection approaches. The latter approaches differ in the applied quality measures for identifying valid reference points and by the number of LU/LC classes (7–12). In an iterative training process, the impact of the chosen PS on a Random Forest image classifier is evaluated. The results are compared to LUCAS reference points that are not pre-processed, which act as a benchmark, and the classification quality is evaluated by independent sets of validation points. The classification results show that the positional correction of LUCAS points has an especially positive effect on the overall classification accuracy. On average, this improves the accuracy by 3.7%. This improvement is lowest for the most rigid sample selection approach, PS₂, and highest for the benchmark data set, PS₀. The highest overall accuracy is 93.1% which is achieved by using the newly developed PS₃; all PS achieve overall accuracies of 80% and higher on average. While the difference in overall accuracy between the PS is likely to be influenced by the respective number of LU/LC classes, we conclude that, overall, LUCAS in-situ data is a suitable source for reference information for large scale high resolution LC mapping using Sentinel-2 imagery. Existing sample selection approaches developed for Landsat imagery can be transferred to Sentinel-2 imagery, achieving comparable semantic accuracies while increasing the spatial resolution. The resulting LC classification product that uses the newly developed PS is available for Germany via DOI: https://doi.org/10.15489/1ccmlap3mn39.     

A comparative assessment between object and pixel-based classification approaches for land use/land cover mapping using SPOT 5 imagery

Abstract

Land use/land cover (LULC) classification with high accuracy is necessary, especially in eco-environment research, urban planning, vegetation condition study and soil management. Over the last decade a number of classification algorithms have been developed for the analysis of remotely sensed data. The most notable algorithms are the object-oriented K-Nearest Neighbour (K-NN), Support Vector Machines (SVMs) and the Decision Trees (DTs) amongst many others. In this study, LULC types of Selangor area were analyzed on the basis of the classification results acquired using the pixel-based and object-based image analysis approaches. SPOT 5 satellite images with four spectral bands from 2003 and 2010 were used to carry out the image classification and ground truth data were collected from Google Earth and field trips. In pixel-based image analysis, a supervised classification was performed using the DT classifier. On the other hand, object-oriented (K-NN) image analysis was evaluated using standard nearest neighbour as classifier. Subsequently SVM object-based classification was performed. Five LULC categories were extracted and the results were compared between them. The overall classification accuracies for 2003 and 2010 showed that the object-oriented (K-NN) (90.5% and 91%) performed better results than the pixel-based DT (68.6% and 68.4%) and object-based SVM (80.6% and 78.15%). In general, the object-oriented (K-NN) performed better than both DTs and SVMs. The obtained LULC classification maps can be used to improve various applications such as change detection, urban design, environmental management and zooning.     

Evaluation of diverse classification approaches for land use/cover mapping in a Mediterranean region utilizing Hyperion data

Information on Earth's land surface cover is commonly obtained through digital image analysis of data acquired from remote sensing sensors. In this study, we evaluated the use of diverse classification techniques in discriminating land use/cover types in a typical Mediterranean setting using Hyperion imagery. For this purpose, the spectral angle mapper (SAM), the object-based and the non-linear spectral unmixing based on artificial neural networks (ANNs) techniques were applied. A further objective had been to investigate the effect of two approaches for training sites selection in the SAM classification, namely of the pixel purity index (PPI) and of the direct selection of training points from the Hyperion imagery assisted by a QuickBird imagery and field-based training sites. Object-based classification outperformed the other techniques with an overall accuracy of 83%. Sub-pixel classification based on the ANN showed an overall accuracy of 52%, very close to that of SAM (48%). SAM applied using the training sites selected directly from the Hyperion imagery supported by the QuickBird image and the field visits returned an increase accuracy by 16%. Yet, all techniques appeared to suffer from the relatively low spatial resolution of the Hyperion imagery, which affected the spectral separation among the land use/cover classes.     

Assessing land use/cover changes: a nationwide multidate spatial database for Mexico

A nationwide multidate GIS database was generated in order to carry out the quantification and spatial characterization of land use/cover changes (LUCC) in Mexico. Existing cartography on land use/cover at a 1:250,000 scale was revised to select compatible inputs regarding the scale, the classification scheme and the mapping method. Digital maps from three different dates (the late 1970s, 1993 and 2000) were revised, evaluated, corrected and integrated into a GIS database. In order to improve the reliability of the database, an attempt was made to assess the accuracy of the digitalisation procedure and to detect and correct unlikely changes due to thematic errors in the maps. Digital maps were overlaid in order to generate LUCC maps, transition matrices and to calculate rates of conversion. Based upon this database, rates of deforestation between 1976 and 2000 were evaluated as 0.25 and 0.76% per year for temperate and tropical forests, respectively.     

Land cover and land use classification performance of machine learning algorithms in a boreal landscape using Sentinel-2 data

ABSTRACT

In 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.     

Spectral angle mapper and object-based classification combined with hyperspectral remote sensing imagery for obtaining land use/cover mapping in a Mediterranean region

In this study, we test the potential of two different classification algorithms, namely the spectral angle mapper (SAM) and object-based classifier for mapping the land use/cover characteristics using a Hyperion imagery. We chose a study region that represents a typical Mediterranean setting in terms of landscape structure, composition and heterogeneous land cover classes. Accuracy assessment of the land cover classes was performed based on the error matrix statistics. Validation points were derived from visual interpretation of multispectral high resolution QuickBird-2 satellite imagery. Results from both the classifiers yielded more than 70% classification accuracy. However, the object-based classification clearly outperformed the SAM by 7.91% overall accuracy (OA) and a relatively high kappa coefficient. Similar results were observed in the classification of the individual classes. Our results highlight the potential of hyperspectral remote sensing data as well as object-based classification approach for mapping heterogeneous land use/cover in a typical Mediterranean setting.     

Stylized environments and ABMs: educational tools for examining the causes and consequences of land use/land cover change

A challenge in land change science is to assess the causes and consequences of LULC change and associated pattern–process relations. Increasingly, land change organizations are examining land use at local to global scales for historical, contemporary and future periods through scenarios that assess population–environment interactions. Spatial analytical tools in GIScience are being used to link people and environment and to search for the distal and proximate factors that affect local to global land use patterns. Spatial simulation models that rely upon complexity theory as the framework and agent-based models as the analytical approach offer the capability to inform through experimentation about land issues important to science and society. Using a stylized landscape where a selected set of key social, geographical and ecological elements are spatially organized, we describe how land dynamics can be examined through agent-based models as educational tools that are useful in the classroom, boardroom and public forums.     

Discrimination of mangrove forests and characterization of adjoining land cover classes using temporal C-band Synthetic Aperture Radar data: A case study of Sundarbans

C-band dual polarization (HH, HV) Synthetic Aperture Radar (SAR) data from Radarsat-2 were used to discriminate and characterize mangrove forests of the Sundarbans. Multi-temporal data acquired during winter and rainy seasons were analysed for the segregation of mangrove forest area. A decision rule based classification involving combination of three-date HH (range −11 to −2 dB) with single-date cross-polarization ratio (2–8) was applied on the datasets for discriminating mangrove forests from other land cover classes. Application of textural measures (entropy and angular second moment) in the aforesaid decision rule based classification produced three broad homogeneous mangrove classes. The area covered by the most homogeneous class increased from January to March and decreased from July to September, and correlated well to the change in the phenological status of the mangroves. Extent of homogeneous areas was more in the eastern region of the Sundarbans than that of the central and western side. Thus, the study revealed that textural measures combined with multi-temporal HH backscatter and single-date cross-polarization ratio in a decision rule classification could be satisfactorily used for characterization of the mangrove forests.     

Choosing pasture maps: An assessment of pasture land classification definitions and a case study of Brazil

Pasture land occupies extensive areas and is increasingly of interest for sustainable intensification, land use diversification, greenhouse gas emission mitigation, and bioenergy expansion. Accurate maps of pasture and other managed land covers are needed for monitoring, intercomparison, assessing potential uses, and planning. Yet, land maps can be generated from different types of classification datasets – i.e. as a land use or land cover type – as well as different sources. In this study our aim was to assess and compare land use and land cover definitions for pasture, and examine variability in the resulting pasture land classification maps. First, we conducted a review of pasture definitions in commonly used mapping databases. We then performed a case study involving Brazil, a dominant global producer of pasture-based livestock. Six geospatial databases were harmonized and compared to each other and to MODIS land cover for Brazil including the Cerrado and Amazon biomes, which are internationally recognized for their ecological value. Total pasture area estimates for Brazil ranged by a factor greater than four, from about 430,000 km² to over 1.7 million km². Our analysis showed high variability in pasture land maps depending on the definitions, methods and underlying datasets used to generate them. The results are illustrative of a symptomatic problem for all manage land datasets, demonstrating the need for land categories studies and geospatial data resources that fully define land terms and describe measurable management attributes. Additionally, the suitability of individual geospatial datasets for different types of land mapping must be better described and reported. These recommendations would help bring more consistency in the consideration of managed lands in research, reporting, and policy development, as demonstrated here for pasture land using six case study datasets from multiple sources.     

Multiple intra-urban land use simulations and driving factors analysis: a case study in Huicheng,China

Simulations of intra-urban land use changes have gradually attracted more attention as these approaches are extremely helpful in regard to decision making and policy formulation. While prior studies mostly focused on methods of developing intra-urban level simulations, very little research has been conducted explain the factors driving intra-urban land use change. Urban planners are highly concerned with how inner-city structures are formed and how they function. Here, to simulate multiple intra-urban land use changes and to identify the contribution of different driving factors, we developed a random forests (RF) algorithm-based cellular automata (CA) simulation model. In this study, the model applied diverse categories of spatial variables, including traffic location factors, environmental factors, public services, and population density, as the driving factors to enhance our understanding of the dynamics of internal urban land use. The CA model was tested using data from the Huicheng district of Huizhou city in the Guangdong province of China. The Model was validated using actual historical land use data from 2000 to 2010. By applying the validated model, multiple intra-urban land use maps were simulated for 2015. Simultaneously, spatial variable importance measures (VIMs) were calculated by using the out-of-bag (OOB) error estimation approach of the RF algorithm. Based on the calculation results, we assessed and analysed the significance of each intra-urban land use driver for this region. This study provides urban planners and relevant scholars with detailed and targeted information that can aid in the formulation of specific planning strategies for different intra-urban land uses and support the future evolution of this area.     

Synergy of airborne LiDAR and Worldview-2 satellite imagery for land cover and habitat mapping: A BIO_SOS-EODHaM case study for the Netherlands

A major challenge is to develop a biodiversity observation system that is cost effective and applicable in any geographic region. Measuring and reliable reporting of trends and changes in biodiversity requires amongst others detailed and accurate land cover and habitat maps in a standard and comparable way. The objective of this paper is to assess the EODHaM (EO Data for Habitat Mapping) classification results for a Dutch case study. The EODHaM system was developed within the BIO_SOS (The BIOdiversity multi-SOurce monitoring System: from Space TO Species) project and contains the decision rules for each land cover and habitat class based on spectral and height information. One of the main findings is that canopy height models, as derived from LiDAR, in combination with very high resolution satellite imagery provides a powerful input for the EODHaM system for the purpose of generic land cover and habitat mapping for any location across the globe. The assessment of the EODHaM classification results based on field data showed an overall accuracy of 74% for the land cover classes as described according to the Food and Agricultural Organization (FAO) Land Cover Classification System (LCCS) taxonomy at level 3, while the overall accuracy was lower (69.0%) for the habitat map based on the General Habitat Category (GHC) system for habitat surveillance and monitoring. A GHC habitat class is determined for each mapping unit on the basis of the composition of the individual life forms and height measurements. The classification showed very good results for forest phanerophytes (FPH) when individual life forms were analyzed in terms of their percentage coverage estimates per mapping unit from the LCCS classification and validated with field surveys. Analysis for shrubby chamaephytes (SCH) showed less accurate results, but might also be due to less accurate field estimates of percentage coverage. Overall, the EODHaM classification results encouraged us to derive the heights of all vegetated objects in the Netherlands from LiDAR data, in preparation for new habitat classifications.     

城市用地空间信息分类与数字重建——以长春百年城市内部用地变化为例

针对单一应用遥感影像难以进行城市内部用地结构分类以及高精度城市内部用地多期空间数据叠置分析中位置误差问题建立了基于"分层分类"与"对象分割"的城市内部用地空间信息数字重建方法。实现对特大城市产业用地(住宅、商业、工业等)以及交通、水系、生态绿地等不同功能结构用地的高精度监测以及历史演变过程的重建。综合集成SPOT5,1︰1万地形图、历史地图及城市规划图等辅助信息对长春城市1905年以来城市用地信息进行分类。研究表明,在专家知识参与下人—机交互解译,集成多源空间信息对实现高精度城市用地空间信息重建具有较高的应用价值,该方法不仅能提高城市用地分类精度而且能提高城市用地空间信息提取效率以及多期空间数据叠置分析的定位精度。     

基于面向对象与规则的Sentinel-2A影像土地覆被分类——以江西省都昌县为例

土地覆被作为地表自然和人工建造物的综合体,是开展土地科学相关研究的重要基础,在遥感大数据背景下,准确、快速、自动化进行土地覆被提取技术一直是遥感研究中的重点。本文基于e Cognition软件,采用面向对象的多尺度分割法,综合考虑地物在遥感影像上的光谱、形状和纹理特征,建立多种地物提取规则。通过模糊函数、支持向量机(SVM)和阈值法对研究区的土地覆被进行分类提取,并与研究区的FROM-GLC10数据和土地利用变更数据进行了对比分析。结果表明:①研究区土地覆被分类的总体精度为97%,Kappa系数为0.96,分类精度较高;②基于10 m分辨率影像,综合使用形状、纹理、光谱信息对于道路的提取具有较好的效果,道路提取Kappa系数为0.84;③分类结果在面积和空间分布上都优于FROM-GLC10数据,与研究区实际土地变更数据保持较好的一致性。基于面向对象与规则的分类方法提取地物能够有效利用多种遥感影像特征,分类精度高,对于处理高分辨率遥感数据具有很好的优势。     

Spatially explicit urban green indicators for characterizing vegetation cover and public green space proximity: a case study on Brussels,Belgium

ABSTRACT

Cities often have a substantial green infrastructure, which provides local ecosystem services that improve the quality of life of urban residents. These services should be explicitly addressed in urban development policies, and areas with insufficient vegetation and limited access to public green spaces should be identified. This paper presents two spatially explicit urban green indicators that are derived using remote sensing imagery, freely available map data and spatial analysis tools from open source geospatial libraries and commercial software. The first indicator represents proportional green cover (public as well as private) in the vicinity of each building within a city. The second indicator quantifies the proximity of public green spaces as walking distances from buildings to actual park entrances. A dasymetric mapping approach was used to take spatial variations in population density into account. This allows representing the indicators from the perspective of citizens instead of buildings, which may be more meaningful for deriving statistics at city level or at the level of neighbourhoods or administrative zones. The potential use of these indicators in a planning context is discussed on a case study carried out for the city of Brussels, Belgium.     

Glacier and glacial lake classification for change detection studies using satellite data: a case study from Baspa basin,western Himalaya

Using high-resolution Google Earth^TM images in conjunction with Landsat images, the glaciers and lakes in the Baspa basin are classified to explore the recent changes. A total number of 109 glaciers (187 ± 3.7 km²) are mapped and subsequently classified as compound valley glaciers, simple valley glaciers, cirques, niches, glacieretes and ice aprons. The compound and simple valley glaciers contribute 67.1 ± 1.3% and 19.8 ± 0.3% to the total glacier cover of the basin. Similarly, a total number of 129 glacial lakes (0.360 ± 0.007 km²) are identified. From 1976 to 2011, the compound valley glaciers have lost a small area of 10.3 ± 0.03% at a rate of 0.41 ± 0.002 km² a^-1, whereas the niche glaciers have lost higher area of 40.1 ± 0.001% at a rate of 0.04 ± 0.0001 km² a^-1. Change detection of two benchmark glacial lakes revealed a progressive expansion during recent decades. The Baspa Bamak proglacial lake has expanded from 0.020 ± 0.0004 km² (2000) to 0.069 ± 0.001 km² (2011). Due to the complete loss of source ice, another glacial lake has expanded from 0.09 ± 0.001 km² (1994) to 0.10 ± 0.002 km² (2011). During the study period, the mean annual temperature that is T_avg, T_min and T_max have increased significantly at the 95% confidence level by 1.5 ^oC (0.070 °C a^-1), 1.8 ^oC (0.076 °C a^-1) and 1.6 ^oC (0.0071 °C a^-1) from 1985 to 2008. However, the precipitation has decreased significantly from 1976 and 1985 to 2008.     

Monitoring surface changes in discontinuous permafrost terrain using small baseline SAR interferometry,object-based classification,and geological features: a case study from Mayo,Yukon Territory,Canada

Permafrost-induced deformation of ground features is threating infrastructure in northern communities. An understanding of permafrost distribution is therefore critical for sustainable adaptation planning and infrastructure maintenance. Considering the large area underlain by permafrost in the Yukon Territory, there is a need for baseline information to characterize the permafrost in this region. In this study, the Differential Interferometric Synthetic Aperture Radar (DInSAR) technique was used to identify areas of ground movement likely caused by changes in permafrost. The DInSAR technique was applied to a series of repeat-pass C-band RADARSAT-2 observations collected in 2015 over the Village of Mayo, in central Yukon Territory, Canada. The conventional DInSAR technique demonstrated that ground deformation could be detected in this area, but the resulting deformation maps contained errors due to a loss of coherence from changes in vegetation and atmospheric phase delay. To address these limitations, the Small BAseline Subset (SBAS) InSAR technique was applied to reduce phase error, thus improving the deformation maps. To understand the relationship between the deformation maps and land cover types, an object-based Random Forest classification was developed to classify the study area into different land cover types. Integration of the InSAR results and the classification map revealed that the built-up class (e.g., airport) was affected by subsidence on the order of ?2 to ?4 cm. The spatial extent of the surface displacement map obtained using the SBAS InSAR technique was then correlated with the surficial geology map. This revealed that much of the main infrastructure in the Village of Mayo is underlain by interbedded glaciofluvial and glaciolacustrine sediments, the latter of which caused the most damage to human made structures. This study provides a method for permafrost monitoring that builds upon the synergistic use of the SBAS InSAR technique, object-based image analysis, and surficial geology data.