We have produced the first 30 m resolution global land-cover maps using Landsat Thematic Mapper (TM) and Enhanced Thematic Mapper Plus (ETM+) data. We have classified over 6600 scenes of Landsat TM data after 2006, and over 2300 scenes of Landsat TM and ETM+ data before 2006, all selected from the green season. These images cover most of the world's land surface except Antarctica and Greenland. Most of these images came from the United States Geological Survey in level L1T (orthorectified). Four classifiers that were freely available were employed, including the conventional maximum likelihood classifier (MLC), J4.8 decision tree classifier, Random Forest (RF) classifier and support vector machine (SVM) classifier. A total of 91,433 training samples were collected by traversing each scene and finding the most representative and homogeneous samples. A total of 38,664 test samples were collected at preset, fixed locations based on a globally systematic unaligned sampling strategy. Two software tools, Global Analyst and Global Mapper developed by extending the functionality of Google Earth, were used in developing the training and test sample databases by referencing the Moderate Resolution Imaging Spectroradiometer enhanced vegetation index (MODIS EVI) time series for 2010 and high resolution images from Google Earth. A unique land-cover classification system was developed that can be crosswalked to the existing United Nations Food and Agriculture Organization (FAO) land-cover classification system as well as the International Geosphere-Biosphere Programme (IGBP) system. Using the four classification algorithms, we obtained the initial set of global land-cover maps. The SVM produced the highest overall classification accuracy (OCA) of 64.9% assessed with our test samples, with RF (59.8%), J4.8 (57.9%), and MLC (53.9%) ranked from the second to the fourth. We also estimated the OCAs using a subset of our test samples (8629) each of which represented a homogeneous area greater than 500 m × 500 m. Using this subset, we found the OCA for the SVM to be 71.5%. As a consistent source for estimating the coverage of global land-cover types in the world, estimation from the test samples shows that only 6.90% of the world is planted for agricultural production. The total area of cropland is 11.51% if unplanted croplands are included. The forests, grasslands, and shrublands cover 28.35%, 13.37%, and 11.49% of the world, respectively. The impervious surface covers only 0.66% of the world. Inland waterbodies, barren lands, and snow and ice cover 3.56%, 16.51%, and 12.81% of the world, respectively.
The family of Kappa indices of agreement claim to compare a map's observed classification accuracy relative to the expected accuracy of baseline maps that can have two types of randomness: (1) random distribution of the quantity of each category and (2) random spatial allocation of the categories. Use of the Kappa indices has become part of the culture in remote sensing and other fields. This article examines five different Kappa indices, some of which were derived by the first author in 2000. We expose the indices' properties mathematically and illustrate their limitations graphically, with emphasis on Kappa's use of randomness as a baseline, and the often-ignored conversion from an observed sample matrix to the estimated population matrix. This article concludes that these Kappa indices are useless, misleading and/or flawed for the practical applications in remote sensing that we have seen. After more than a decade of working with these indices, we recommend that the profession abandon the use of Kappa indices for purposes of accuracy assessment and map comparison, and instead summarize the cross-tabulation matrix with two much simpler summary parameters: quantity disagreement and allocation disagreement. This article shows how to compute these two parameters using examples taken from peer-reviewed literature.
The Defense Meteorological Satellite Program/Operational Linescane System (DMSP/OLS) stable night-time light (NTL) data showed great potential in urban extent mapping across a variety of scales with historical records dating back to 1990s. In order to advance this data, a systematic methodology review on NTL-based urban extent mapping was carried out, with emphases on four aspects including the saturation of luminosity, the blooming effect, the intercalibration of time series, and their temporal pattern adjustment. We think ancillary features (e.g. land surface conditions and socioeconomic activities) can help reveal more spatial details in urban core regions with high digital number (DN) values. In addition, dynamic optimal thresholds are needed to address issues of different exaggeration of NTL data in the large scale urban mapping. Then, we reviewed three key aspects (reference region, reference satellite/year, and calibration model) in the current intercalibration framework of NTL time series, and summarized major reference regions in literature that were used for intercalibration, which is critical to achieve a globally consistent series of NTL DN values over years. Moreover, adjustment of temporal pattern on intercalibrated NTL series is needed to trace the urban sprawl process, particularly in rapidly developing regions. In addition, we analysed those applications for urban extent mapping based on the new generation NTL data of Visible/Infrared Imager/Radiometer Suite. Finally, we prospected the challenges and opportunities including the improvement of temporally inconsistent NTL series, mitigation of spatial heterogeneity of blooming effect in NTL, and synthesis of different NTL satellites, in global urban extent mapping.
As an intrinsic property of natural materials, land surface emissivity (LSE) is an important surface parameter and can be derived from the emitted radiance measured from space. Besides radiometric calibration and cloud detection, two main problems need to be resolved to obtain LSE values from space measurements. These problems are often referred to as land surface temperature (LST) and emissivity separation from radiance at ground level and as atmospheric corrections in the literature. To date, many LSE retrieval methods have been proposed with the same goal but different application conditions, advantages, and limitations. The aim of this article is to review these LSE retrieval methods and to provide technical assistance for estimating LSE from space. This article first gives a description of the theoretical basis of LSE measurements and then reviews the published methods. For clarity, we categorize these methods into (1) (semi-)empirical or theoretical methods, (2) multi-channel temperature emissivity separation (TES) methods, and (3) physically based methods (PBMs). This article also discusses the validation methods, which are of importance in verifying the uncertainty and accuracy of retrieved emissivity. Finally, the prospects for further developments are given.
Characterizations of land-cover dynamics are among the most important applications of Earth observation data, providing insights into management, policy and science. Recent progress in remote sensing and associated digital image processing offers unprecedented opportunities to detect changes in land cover more accurately over increasingly large areas, with diminishing costs and processing time. The advent of high-spatial-resolution remote-sensing imagery further provides opportunities to apply change detection with object-based image analysis (OBIA), that is, object-based change detection (OBCD). When compared with the traditional pixel-based change paradigm, OBCD has the ability to improve the identification of changes for the geographic entities found over a given landscape. In this article, we present an overview of the main issues in change detection, followed by the motivations for using OBCD as compared to pixel-based approaches. We also discuss the challenges caused by the use of objects in change detection and provide a conceptual overview of solutions, which are followed by a detailed review of current OBCD algorithms. In particular, OBCD offers unique approaches and methods for exploiting high-spatial-resolution imagery, to capture meaningful detailed change information in a systematic and repeatable manner, corresponding to a wide range of information needs.
The Visible Infrared Imaging Radiometer Suite (VIIRS) Day/Night Band (DNB) collects global low-light imaging data that have significant improvements over comparable data collected for 40 years by the DMSP Operational Linescan System. One of the prominent features of DNB data is the detection of electric lighting present on the Earth's surface. Most of these lights are from human settlements. VIIRS collects source data that could be used to generate monthly and annual science grade global radiance maps of human settlements with electric lighting. There are a substantial number of steps involved in producing a product that has been cleaned to exclude background noise, solar and lunar contamination, data degraded by cloud cover, and features unrelated to electric lighting (e.g. fires, flares, volcanoes). This article describes the algorithms developed for the production of high-quality global VIIRS night-time lights. There is a broad base of science users for VIIRS night-time lights products, ranging from land-use scientists, urban geographers, ecologists, carbon modellers, astronomers, demographers, economists, and social scientists.
Machine learning offers the potential for effective and efficient classification of remotely sensed imagery. The strengths of machine learning include the capacity to handle data of high dimensionality and to map classes with very complex characteristics. Nevertheless, implementing a machine-learning classification is not straightforward, and the literature provides conflicting advice regarding many key issues. This article therefore provides an overview of machine learning from an applied perspective. We focus on the relatively mature methods of support vector machines, single decision trees (DTs), Random Forests, boosted DTs, artificial neural networks, and k-nearest neighbours (k-NN). Issues considered include the choice of algorithm, training data requirements, user-defined parameter selection and optimization, feature space impacts and reduction, and computational costs. We illustrate these issues through applying machine-learning classification to two publically available remotely sensed data sets.
Unmanned aerial vehicles (UAVs) or remotely piloted aircraft systems are new platforms that have been increasingly used over the last decade in Europe to collect data for forest research, thanks to the miniaturization and cost reduction of GPS receivers, inertial navigation system, computers, and, most of all, sensors for remote sensing. In this review, after describing the regulatory framework for the operation of UAVs in the European Union (EU), an overview of applications in forest research is presented, followed by a discussion of the results obtained from the analysis of different case studies. Rotary-wing and fixed-wing UAVs are equally distributed among the case studies, while ready-to-fly solutions are preferred over self-designed and developed UAVs. Most adopted technologies are visible-red, green, and blue, multispectral in visible and near-infrared, middle-infrared, thermal infrared imagery, and lidar. The majority of current UAV-based applications for forest research aim to inventory resources, map diseases, classify species, monitor fire and its effects, quantify spatial gaps, and estimate post-harvest soil displacement. Successful implementation of UAVs in forestry depends on UAV features, such as flexibility of use in flight planning, low cost, reliability and autonomy, and capability of timely provision of high-resolution data. Unfortunately, the fragmented regulations among EU countries, a result of the lack of common rules for operating UAVs in Europe, limit the chance to operate within Europe's boundaries and prevent research mobility and exchange opportunities. Nevertheless, the applications of UAVs are expanding in different domains, and the use of UAVs in forestry will increase, possibly leading to a regular utilization for small-scale monitoring purposes in Europe when recent technologies (i.e. hyperspectral imagery and lidar) and methodological approaches will be consolidated.
Image classification is a complex process that may be affected by many factors. This paper examines current practices, problems, and prospects of image classification. The emphasis is placed on the summarization of major advanced classification approaches and the techniques used for improving classification accuracy. In addition, some important issues affecting classification performance are discussed. This literature review suggests that designing a suitable image-processing procedure is a prerequisite for a successful classification of remotely sensed data into a thematic map. Effective use of multiple features of remotely sensed data and the selection of a suitable classification method are especially significant for improving classification accuracy. Non-parametric classifiers such as neural network, decision tree classifier, and knowledge-based classification have increasingly become important approaches for multisource data classification. Integration of remote sensing, geographical information systems (GIS), and expert system emerges as a new research frontier. More research, however, is needed to identify and reduce uncertainties in the image-processing chain to improve classification accuracy.
The normalized difference water index (NDWI) of McFeeters (1996) was modified by substitution of a middle infrared band such as Landsat TM band 5 for the near infrared band used in the NDWI. The modified NDWI (MNDWI) can enhance open water features while efficiently suppressing and even removing built-up land noise as well as vegetation and soil noise. The enhanced water information using the NDWI is often mixed with built-up land noise and the area of extracted water is thus overestimated. Accordingly, the MNDWI is more suitable for enhancing and extracting water information for a water region with a background dominated by built-up land areas because of its advantage in reducing and even removing built-up land noise over the NDWI.,The normalized difference water index (NDWI) of McFeeters ( 1996 ) was modified by substitution of a middle infrared band such as Landsat TM band 5 for the near infrared band used in the NDWI. The modified NDWI (MNDWI) can enhance open water features while efficiently suppressing and even removing built-up land noise as well as vegetation and soil noise. The enhanced water information using the NDWI is often mixed with built-up land noise and the area of extracted water is thus overestimated. Accordingly, the MNDWI is more suitable for enhancing and extracting water information for a water region with a background dominated by built-up land areas because of its advantage in reducing and even removing built-up land noise over the NDWI.
Extended attribute profiles and extended multi-attribute profiles are presented for the analysis of hyperspectral high-resolution images. These extended profiles are based on morphological attribute filters and, through a multi-level analysis, are capable of extracting spatial features that can better model the spatial information, with respect to conventional extended morphological profiles. The features extracted by the proposed extended profiles were considered for a classification task. Two hyperspectral high-resolution datasets acquired for the city of Pavia, Italy, were considered in the analysis. The effectiveness of the introduced operators in modelling the spatial information was proved by the higher classification accuracies obtained with respect to those achieved by a conventional extended morphological profile.
Daily daytime Advanced Very High Resolution Radiometer (AVHRR) 4-km global area coverage data have been processed to produce a Normalized Difference Vegetation Index (NDVI) 8-km equal-area dataset from July 1981 through December 2004 for all continents except Antarctica. New features of this dataset include bimonthly composites, NOAA-9 descending node data from August 1994 to January 1995, volcanic stratospheric aerosol correction for 1982-1984 and 1991-1993, NDVI normalization using empirical mode decomposition/reconstruction to minimize varying solar zenith angle effects introduced by orbital drift, inclusion of data from NOAA-16 for 2000-2003 and NOAA-17 for 2003-2004, and a similar dynamic range with the MODIS NDVI. Two NDVI compositing intervals have been produced: a bimonthly global dataset and a 10-day Africa-only dataset. Post-processing review corrected the majority of dropped scan lines, navigation errors, data drop outs, edge-of-orbit composite discontinuities, and other artefacts in the composite NDVI data. All data are available from the University of Maryland Global Land Cover Facility ( http://glcf.umiacs.umd.edu/data/gimms/ ).
The empirical approach of remote sensing has a proven capability to provide timely and accurate information on inland and near-coastal transitional waters. This article gives a thorough review of empirical algorithms for quantitatively estimating a variety of parameters from space-borne, airborne and in situ remote sensors in inland and transitional waters, including chlorophyll-a, total suspended solids, Secchi disk depth (z SD ), turbidity, absorption by coloured dissolved organic matter (a CDOM ) and other parameters, for example, phycocyanin. Current remote-sensing instruments are also reviewed. The theoretical basis of the empirical algorithms is given using fundamental bio-optical theory of the inherent optical properties (IOPs). Bands, band ratios and band arithmetic algorithms that could be used to produce common biogeophysical products for inland/transitional waters are identified. The article discusses the potential role that empirical algorithms could play alongside more advanced model-based algorithms in the future of water remote sensing, especially for near real-time operational monitoring systems. The article aims to describe the current status of empirical remote sensing in inland and near-coastal transitional waters and provide a useful reference to workers. It does not cover 'inversion' algorithms.
This article presents an automated Sentinel-1-based processing chain designed for flood detection and monitoring in near-real-time (NRT). Since no user intervention is required at any stage of the flood mapping procedure, the processing chain allows deriving time-critical disaster information in less than 45 min after a new data set is available on the Sentinel Data Hub of the European Space Agency (ESA). Due to the systematic acquisition strategy and high repetition rate of Sentinel-1, the processing chain can be set up as a web-based service that regularly informs users about the current flood conditions in a given area of interest. The thematic accuracy of the thematic processor has been assessed for two test sites of a flood situation at the border between Greece and Turkey with encouraging overall accuracies between 94.0% and 96.1% and Cohen's kappa coefficients (κ) ranging from 0.879 to 0.910. The accuracy assessment, which was performed separately for the standard polarizations (VV/VH) of the interferometric wide swath (IW) mode of Sentinel-1, further indicates that under calm wind conditions, slightly higher thematic accuracies can be achieved by using VV instead of VH polarization data.
Advances in computer vision and the parallel development of unmanned aerial vehicles (UAVs) allow for the extensive use of UAV in forest inventory and in indirect measurements of tree features. We used UAV-sensed high-resolution imagery through photogrammetry and Structure from Motion (SfM) to estimate tree heights and crown diameters. We reconstructed 3D structures from 2D image sequences for two study areas (25 × 25 m). Species composition for Plot 1 included Norway spruce (Picea abies L.) together with European larch (Larix decidua Mill.) and Scots pine (Pinus sylvestris L.), whereas Plot 2 was mainly Norway spruce and Scots pine together with scattered individuals of European larch and Silver birch (Betula pendula Roth.). The involved workflow used canopy height models (CHMs) for the extraction of height, the smoothing of raster images for the determination of the local maxima, and Inverse Watershed Segmentation (IWS) for the estimation of the crown diameters with the help of a geographical information system (GIS). Finally, we validated the accuracies of the two methods by comparing the UAV results with ground measurements. The results showed higher agreement between field and remote-sensed data for heights than for crown diameters based on RMSE%, which were in the range 11.42-12.62 for height and 14.29-18.56 for crown diameter. Overall, the accuracy of the results was acceptable and showed that the methods were feasible for detecting tree heights and crown diameter.
The use of satellite remote sensing for the mapping of snow-cover characteristics has a long-lasting history reaching back until the 1960s. Because snow cover plays an important role in the Earth's climate system, it is necessary to map snow-cover extent and snow mass in both high temporal and high spatial resolutions. This task can only be achieved by the use of remotely sensed data. Many different sensors have been used in the past decades with various algorithms and respective accuracies. This article provides an overview of the most common methods. The limitations, advantages and drawbacks will be illustrated while error sources and strategies on how to ease their impact will be reviewed. Beginning with a short summary of the physical and spectral properties of snow, methods to map snow extent from the reflective part of the spectrum, algorithms to estimate snow water equivalent (SWE) from passive microwave (PM) data and the combination of both spectra will be delineated. At the end, the reader should have an overarching overview of what is currently possible and the difficulties that can occur in the context of snow-cover mapping from the reflective and microwave parts of the spectrum.
In recent years, satellite imagery has greatly improved in both spatial and spectral resolution. One of the major unsolved problems in highly developed remote sensing imagery is the manual selection and combination of appropriate features according to spectral and spatial properties. Deep learning framework can learn global and robust features from the training data set automatically, and it has achieved state-of-the-art classification accuracies over different image classification tasks. In this study, a technique is proposed which attempts to classify hyperspectral imagery by incorporating deep learning features. Firstly, deep learning features are extracted by multiscale convolutional auto-encoder. Then, based on the learned deep learning features, a logistic regression classifier is trained for classification. Finally, parameters of deep learning framework are analysed and the potential development is introduced. Experiments are conducted on the well-known Pavia data set which is acquired by the reflective optics system imaging spectrometer sensor. It is found that the deep learning-based method provides a more accurate classification result than the traditional ones.
Research on global environmental change requires new data processing and analysis tools that can integrate heterogeneous geospatial data from real-time in situ measurement, remote sensing (RS) and geographic information systems (GISs) at the global scale. The rapid growth of virtual globes for global geospatial information management and display holds promise to meet such a requirement. Virtual globes, Google Earth in particular, enable scientists around the world to communicate their data and research findings in an intuitive three-dimensional (3D) global perspective. Different from traditional GIS, virtual globes are low cost and easy to use in data collection, exploration and visualization. Since 2005, a considerable number of papers have been published in peer-reviewed journals and proceedings from a variety of disciplines. In this review, we examine the development and applications of Google Earth and highlight its merits and limitations for Earth science studies at the global scale. Most limitations are not unique to Google Earth, but to all virtual globe products. Several recent efforts to increase the functionalities in virtual globes for studies at the global scale are introduced. The power of virtual globes in their current generations is mostly restricted to functions as a 'geobrowser'; a better virtual globe tool for Earth science and global environmental change studies is described.
The aim of this study is twofold: first, to present a survey of the actual and most advanced methods related to the use of unmanned aerial systems (UASs) that emerged in the past few years due to the technological advancements that allowed the miniaturization of components, leading to the availability of small-sized unmanned aerial vehicles (UAVs) equipped with Global Navigation Satellite Systems (GNSS) and high quality and cost-effective sensors; second, to advice the target audience - mostly farmers and foresters - how to choose the appropriate UAV and imaging sensor, as well as suitable approaches to get the expected and needed results of using technological tools to extract valuable information about agroforestry systems and its dynamics, according to their parcels' size and crop's types.Following this goal, this work goes beyond a survey regarding UAS and their applications, already made by several authors. It also provides recommendations on how to choose both the best sensor and UAV, in according with the required application. Moreover, it presents what can be done with the acquired sensors' data through theuse of methods, procedures, algorithms and arithmetic operations. Finally, some recent applications in the agroforestry research area are presented, regarding the main goal of each analysed studies, the used UAV, sensors, and the data processing stage to reach conclusions.
Accurate estimates of papyrus (Cyperus papyrus) biomass are critical for an efficient papyrus swamp monitoring and management system. The objective of this study was to test the utility of random forest (RF) regression and two narrow-band vegetation indices in estimating above-ground biomass (AGB) for complex and densely vegetated swamp canopies. The normalized difference vegetation index (NDVI) and enhanced vegetation index (EVI) were calculated from field spectrometry data and fresh AGB was measured in 82 quadrats at three different areas in the iSimangaliso Wetland Park, South Africa. NDVI was calculated from all possible band combinations of the electromagnetic spectrum (350 and 2500 nm), while EVI was calculated from possible band combinations in the blue, red, and near infrared of the spectrum. Backward feature elimination and RF regression were used as variable selection and modelling techniques to predict papyrus AGB. Results showed that the effective portions of electromagnetic spectrum for estimation AGB of papyrus swamp were located within the blue, red, red-edge, and near-infrared regions. The three best selected EVIs were computed from bands located at (i) 445, 682, and 829 nm, (ii) 497, 676, and 1091 nm, and (iii) 495, 678, and 1120 nm. These indices produced better predictive accuracies (R 2 = 0.90; root mean square error of prediction (RMSEP) = 0.289 kg m −2 ; 7.99% of the mean) than the best selected NDVIs (R 2 = 0.85; RMSEP = 0.343 kgm −2 ; 9.49% of the mean) that were calculated from bands located at (i) 739 and 829 nm, (ii) 739 and 814 nm, (iii) 744 and 789 nm, and (iv) 734 and 909 nm. The results of the present study demonstrate the utility of narrow-band vegetation indices (EVI and NDVI) and RF regression in estimating papyrus AGB at high density, a previously challenging task with broadband satellite sensors.