Showing papers in "Remote Sensing Letters in 2015"
TL;DR: Comparative experiments conducted over widely used hyperspectral data indicate that DCNNs-LR classifier built in this proposed deep learning framework provides better classification accuracy than previous hyperspectRAL classification methods.
Abstract: In this letter, a novel deep learning framework for hyperspectral image classification using both spectral and spatial features is presented. The framework is a hybrid of principal component analysis, deep convolutional neural networks (DCNNs) and logistic regression (LR). The DCNNs for hierarchically extract deep features is introduced into hyperspectral image classification for the first time. The proposed technique consists of two steps. First, feature map generation algorithm is presented to generate the spectral and spatial feature maps. Second, the DCNNs-LR classifier is trained to get useful high-level features and to fine-tune the whole model. Comparative experiments conducted over widely used hyperspectral data indicate that DCNNs-LR classifier built in this proposed deep learning framework provides better classification accuracy than previous hyperspectral classification methods.
422 citations
TL;DR: Comparative experiments demonstrate that the uses of waveform representation and deep Boltzmann machines contribute to the improvement of classification accuracies of tree species.
Abstract: Our work addresses the problem of extracting and classifying tree species from mobile LiDAR data. The work includes tree preprocessing and tree classification. In tree preprocessing, voxel-based upward-growing filtering is proposed to remove ground points from the mobile LiDAR data, followed by a tree segmentation that extracts individual trees via Euclidean distance clustering and voxel-based normalized cut segmentation. In tree classification, first, a waveform representation is developed to model geometric structures of trees. Then, deep learning techniques are used to generate high-level feature abstractions of the trees’ waveform representations. Quantitative analysis shows that our algorithm achieves an overall accuracy of 86.1% and a kappa coefficient of 0.8 in classifying urban tree species using mobile LiDAR data. Comparative experiments demonstrate that the uses of waveform representation and deep Boltzmann machines contribute to the improvement of classification accuracies of tree species.
155 citations
TL;DR: An unsupervised feature learning framework based on auto-encoder is proposed to learn sparse feature representations for remote-sensing imagery retrieval that is comparable or superior to that of the state-of-the-art method.
Abstract: An unsupervised feature learning framework based on auto-encoder is proposed to learn sparse feature representations for remote-sensing imagery retrieval in this letter. The low-level feature descriptors are extracted and exploited to learn a set of feature extractors, which are then used to encode the low-level feature descriptors to generate new sparse features. The learned feature representations are applied to aerial images randomly selected from the University of California Merced data set. The results indicate that the performance of our proposed framework is comparable or superior to that of the state-of-the-art method. The framework is proved to be an effective approach to manage the huge volume of remote-sensing data and to retrieve the desired remote-sensing imagery.
76 citations
TL;DR: In this paper, the authors investigated the bias and the error variance of the AHI infrared measurements under clear-sky conditions by using a numerical weather prediction model and a radiative transfer model.
Abstract: The Japan Himawari-8 geostationary satellite was successfully launched into the geosynchronous orbit around 140°E on 17 October 2014. The Advanced Himawari Imager (AHI) onboard the Himawari-8 has its channels 7–16 covering from the short- to the thermal-infrared bands, of which observations can be assimilated into a data assimilation system to improve the atmospheric analysis. Before conducting any AHI data assimilation experiments, it is the first and critical step to correctly quantify AHI bias and error variance, since these two variables are required in a data assimilation system. This study investigates the bias and the error variance of the AHI infrared measurements under clear-sky conditions by using a numerical weather prediction model and a radiative transfer model. Overall, AHI observations agree favourably with the model simulations. It is noted that channels 7–14 has a cold bias of approximately 1.0 K while the cold bias reaches around 2.0 ~ 3.0 K for the longwave Channel 15 and Channel 16. Th...
62 citations
TL;DR: Performance and effectiveness of a novel ensemble learning algorithm, rotation forest (RotFor), aiming to build diverse and accurate classifiers, was investigated for the first time in object-based classification using a WorldView-2 (WV-2) satellite image and confirmed that integration of ancillary data increased the classification accuracy.
Abstract: Machine learning algorithms reported to be robust and superior to the conventional parametric classifiers have been recently employed in object-based classification. Within these algorithms, ensemble learning methods that construct set of individual classifiers and combining their predictions to make final decision about unlabelled data have been successfully applied. In this study, performance and effectiveness of a novel ensemble learning algorithm, rotation forest (RotFor) aiming to build diverse and accurate classifiers, was investigated for the first time in object-based classification using a WorldView-2 (WV-2) satellite image. Also, the combination of satellite imagery and ancillary data (i.e. normalized difference vegetation index and principal components) were assessed. Random forest (RF), support vector machine (SVM) and nearest neighbour (NN) algorithms were also used as benchmark classifiers to evaluate the power of RotFor. The classification results confirmed that integration of ancillary dat...
55 citations
TL;DR: In this paper, the authors presented an approach using dynamic threshold techniques and utilizing time series to generate a data set containing detected surface water bodies on a global scale with daily temporal resolution.
Abstract: The understanding and assessment of surface water variability of inland water bodies, for example, due to climate variability and human impact, requires steady and continuous information about its inter- and intra-annual dynamics. In this letter, we present an approach using dynamic threshold techniques and utilizing time series to generate a data set containing detected surface water bodies on a global scale with daily temporal resolution. Exemplary results for the year 2013 that were based on moderate resolution imaging spectroradiometer products are presented in this letter. The main input data sets for the presented product were MOD09GQ/MYD09GQ and MOD10A1/MYD10A1 with a spatial resolution of 250 m and 500 m, respectively. Using the static water mask MOD44W, we extracted training pixels to generate dynamic thresholds for individual data sets on daily basis. In a second processing step, the generated sequences of water masks were utilized to interpolate the results for any missing observations, either ...
52 citations
TL;DR: Results based on the root mean square errors and coefficient of determination (R2) show that artificial neural networks outperform the inversion model and the regression tree.
Abstract: Neural networks are widely used as predictors in several fields of applications, such as prediction of shallow water depth. The purpose of this study is to investigate the performance of two artificial neural networks models as potential methods in bathymetry. A comparison approach is used to evaluate network models, the regression tree and an inversion model. The high-resolution IKONOS and moderate-resolution Landsat satellite images serve as the case studies, and results based on the root mean square errors and coefficient of determination (R2) show that artificial neural networks outperform the inversion model and the regression tree.
44 citations
TL;DR: In this article, a Modified Histogram method was proposed and applied to its short-wave infrared (SWIR) band to estimate the water fraction within each single pixel, and the estimated surface water fraction was then compared with that derived from a corresponding Moderate Resolution Imaging Spectroradiometer (MODIS) image.
Abstract: Measuring surface water using remote sensing technology is an essential research topic in many research areas, including flood-related studies and water resource management. Recent advances in satellite remote sensing provide more efficient ways of monitoring surface water from space. As a newly available data source with a high frequency of global coverage and moderate resolution, Visible Infrared Imaging Radiometer Suite on board the Suomi National Polar-orbiting Partnership (Suomi NPP-VIIRS) has the advantage of monitoring the earth surface intensively and continuously. This study conducts an exploratory evaluation on the performance of Suomi NPP-VIIRS data for surface water detection. A Modified Histogram method was proposed and applied to its shortwave infrared (SWIR) band to estimate the water fraction within each single pixel. The estimated surface water fraction was then compared with that derived from a corresponding Moderate Resolution Imaging Spectroradiometer (MODIS) image and evaluated using ...
44 citations
TL;DR: In this article, the authors proposed a method using Landsat-8 Operational Land Imager and Thermal Infrared Sensor data for agricultural plastic cover detection using four normalized difference indices: the green Normalized Difference Vegetation Index and three ad hoc spectral indices purposely created for this study.
Abstract: In this article, we are proposing a method using Landsat-8 Operational Land Imager and Thermal Infrared Sensor data for agricultural plastic cover detection. Four normalized difference indices were combined in the procedure described to achieve consistent results: the green Normalized Difference Vegetation Index and three ad hoc spectral indices purposely created for this study (rescaled brightness temperature, Plastic Surface Index and Normalized Difference Sandy Index). The sampling time related to the preliminary collection of spectral information on plastic surfaces was reduced using information gathered through the Quality Assessment and Cloud Quality bands. The overall accuracies observed were on average higher than 80%,and the low cost of the open data set used, lacking ancillary data, demonstrated the reliability of the proposed method, proving its suitability for environmental and agricultural monitoring over large areas.
43 citations
TL;DR: The Global Snowpack as discussed by the authors is a set of global snow cover parameters for the first time in medium resolution for the full globe and without the compromising effects of cloud coverage or polar darkness.
Abstract: With the Global SnowPack, we present a set of global snow cover parameters – for the first time in medium resolution for the full globe and without the compromising effects of cloud coverage or polar darkness. Over 1.2 million single data sets were processed to prepare the Global SnowPack between September 2000 and 2015 – with around 246 more being added every day. Snow cover duration (SCD), early and late season SCD, and statistical products are the main components of the Global SnowPack which can be used to analyse shifts and trends of global snow cover characteristics as well as extreme events. The 500 m resolution allows for applications on a subcatchment scale. One example for a possible application is included, focusing on a detailed view of the California and Volga Basin snow cover characteristics. The Global SnowPack reveals areas with extremely low SCD in 2013/2014 and 2014/2015 which is one reason for the severe droughts in California.
42 citations
TL;DR: In this paper, a simple and effective phenology-based algorithm was developed to detect and map rubber tree plantations in Xishuangbanna, a prefecture in southwest China's Yunnan province.
Abstract: A simple and effective phenology-based algorithm was developed to detect and map rubber tree plantations in Xishuangbanna, a prefecture in southwest China’s Yunnan province. This algorithm highlighted the unique phenological characteristics of deciduous rubber tree plantations during the dry season. Phenology of rubber tree plantations and natural evergreen forests was delineated with Normalized Difference Vegetation Index (NDVI), Land Surface Water Index (LSWI) and Normalized Burn Ratio (NBR) derived from Landsat Thematic Mapper (TM), Enhanced Thematic Mapper plus (ETM+) and Operational Land Imager imagery during 2009–2014. The results showed that the differences of NBR were larger than those of NDVI and LSWI from defoliation stage to foliation stage. Then, the change rate of NBR derived between defoliation stage and foliation stage was used to map rubber tree plantations in 2014, by combining a Landsat-based forest mask and a Digital Elevation Model mask. Our study demonstrates that Landsat imagery hold...
TL;DR: In this article, a set of algorithms for satellite mapping of green leaf area index (LAI) in C3 and C4 crops were developed using in situ hyperspectral reflectance and green LAI data, collected across eight years (2001-2008) at three AmeriFlux sites in Nebraska USA over irrigated and rain-fed maize and soybean.
Abstract: This study developed a set of algorithms for satellite mapping of green leaf area index (LAI) in C3 and C4 crops. In situ hyperspectral reflectance and green LAI data, collected across eight years (2001–2008) at three AmeriFlux sites in Nebraska USA over irrigated and rain-fed maize and soybean, were used for algorithm development. The hyperspectral reflectance was resampled to simulate the spectral bands of sensors aboard operational satellites (Aqua and Terra: MODIS, Landsat: TM/ETM+), a legacy satellite (Envisat: MERIS), and future satellites (Sentinel-2, Sentinel-3, and Venµs). Among 15 vegetation indices (VIs) examined, five VIs – wide dynamic range vegetation index (WDRVI), green WDRVI, red edge WDRVI, and green and red edge chlorophyll indices – had a minimal noise equivalent for estimating maize and soybean green LAI ranging from 0 to 6.5 m2 m−2. The algorithms were validated using MODIS, TM/ETM+, and MERIS satellite data. The root mean square error of green LAI prediction in both crops from all s...
TL;DR: In this letter, a new pixel-wise learning method based on deep belief networks (DBNs) for object recognition is proposed and the recognition results demonstrate the accuracy and efficiency of the proposed method.
Abstract: Object recognition has been one of the hottest issues in the field of remote sensing image analysis. In this letter, a new pixel-wise learning method based on deep belief networks (DBNs) for object recognition is proposed. The method is divided into two stages, the unsupervised pre-training stage and the supervised fine-tuning stage. Given a training set of images, a pixel-wise unsupervised feature learning algorithm is utilized to train a mixed structural sparse restricted Boltzmann machine (RBM). After that, the outputs of this RBM are put into the next RBM as inputs. By stacking several layers of RBM, the deep generative model of DBNs is built. At the fine-tuning stage, a supervised layer is attached to the top of the DBN and labels of the data are put into this layer. The whole network is then trained using the back-propagation (BP) algorithm with sparse penalty. Finally, the deep model generates good joint distribution of images and their labels. Comparative experiments are conducted on our dataset a...
TL;DR: In this article, the authors investigated the potential for using broadband multispectral vegetation indices to detect impacts of oil pollution on vegetation conditions using data acquired at the visible, near infrared and shortwave infrared wavelengths.
Abstract: Vegetation health and vigour may be affected by oil leakage or pollution. This effect can alter a plant’s behaviour and may be used as evidence for detecting oil pollution in the environment. Satellite remote sensing has been shown to be an effective tool and approach to detect and monitor vegetation health and status in polluted areas. Previous research has used vegetation indices derived from remotely sensed satellite data to monitor vegetation health. This study investigated the potential for using broadband multispectral vegetation indices to detect impacts of oil pollution on vegetation conditions. Twenty indices were explored and evaluated in this study. The indices use data acquired at the visible, near infrared and shortwave infrared wavelengths. Comparative index values from the 37 oil polluted and non-polluted (control) sites show that 12 Broadband multispectral vegetation indices (BMVIs) indicated significant differences (p-value < 0.05) between pre- and post-spill observations. The 12 BMVI val...
TL;DR: In this paper, a conventional digital camera with a charge coupled device was integrated into an UAV to capture digital aerial images of paddy rice (Oryza sativa L.) at an altitude of 50m.
Abstract: With the rise of large-scale crop plantation in China, inexpensive but efficient remote-sensing measures for predicting the nitrogen status of crops are needed for optimal fertilizer management. In this research, a conventional digital camera with a charge coupled device was integrated into an unmanned aerial vehicle (UAV) to capture digital aerial images of paddy rice (Oryza sativa L.) at an altitude of 50 m. The fluorescence emissions of the rice leaves under light excitation were used by Multiplex® to non-destructively assess the chlorophyll and polyphenol content. The nitrogen balance index (NBI) of the rice leaves, known as the ratio of chlorophyll to polyphenols, was used to accurately determine canopy nitrogen concentrations. The dark green colour index (DGCI) available from the aerial images was used to assess the nitrogen concentrations in the field. It was found that DGCI values predicted the nitrogen concentrations and NBI with R2 (coefficient of determination) = 0.672 (p < 0.001) and R2 = 0.71...
TL;DR: In this article, the authors demonstrate and evaluate the use and potential advantages of AUV field data collection for calibration and validation of seagrass habitat mapping of shallow waters (< 5 m), from multispectral satellite imagery.
Abstract: Repeatable and accurate seagrass mapping is required for understanding seagrass ecology and supporting management decisions. For shallow ( 20 m). Here we demonstrate and evaluate the use and potential advantages of AUV field data collection for calibration and validation of seagrass habitat mapping of shallow waters (< 5 m), from multispectral satellite imagery. The study was conducted in the seagrass habitats of the Eastern Banks (142 km2), Moreton Bay, Australia. In the field, georeferenced photos of the seagrass were collected along transects via snorkelling or an AUV. Photos from both collection methods were analysed manually for seagrass species composition and then used as calibration and validation data to map seagrass using an established semi-automated object based mapping routine. A comparison of the relative advantages and disadvantages of AUV and snorkeller collected field data sets and their influence on the mapping routine was conducted. AUV data collection was more consistent, repeatable and safer in comparison to snorkeller transects. Inclusion of deeper water AUV data resulted in mapping of a larger extent of seagrass (~7 km2, 5 % of study area) in the deeper waters of the site. Although overall map accuracies did not differ considerably, inclusion of the AUV data from deeper water transects corrected errors in seagrass mapped at depths to 5 m, but where the bottom is visible on satellite imagery. Our results demonstrate that further development of AUV technology is justified for the monitoring of seagrass habitats in ongoing management programs.
TL;DR: In this article, the authors demonstrate an approach to downscaling air temperatures for site-level studies using airborne LiDAR data and remote microclimate loggers. And they recommend that researchers consider the scales relevant to specific applications when using their approach to develop site-specific spatio-temporal models.
Abstract: A spatial mismatch exists between regional climate models and conditions experienced by individual organisms. We demonstrate an approach to downscaling air temperatures for site-level studies using airborne LiDAR data and remote microclimate loggers. In 2012–2013, we established a temperature logger network in the forested region of central Missouri, USA, and obtained sub-hourly meteorological measurements from a centrally located weather station. We then used linear mixed models within an information theoretic approach to evaluate hourly and seasonal effects of insolation, vegetation structure, elevation, and meteorological measurements on near-surface air temperatures. The best-supported models predicted fine-scale temperatures with high accuracy during both the winter and growing seasons. We recommend that researchers consider the scales relevant to specific applications when using our approach to develop site-specific spatio-temporal models.
TL;DR: Describing of types of bio-optical algorithm is present as well as a procedure to define the most suitable terminology, based on the goal, processes and products of the bio-Optical algorithm.
Abstract: Bio-optical algorithms have been classify with different terms such as empirical, semi-empirical, semi-analytical, quasi-analytical or analytical algorithms. However, one algorithm has been classified differently in remote sensing literature and a lack of a consistent terminology was found. In this article, description of types of bio-optical algorithm is present as well as a procedure to define the most suitable terminology. This procedure is based on the goal, processes and products of the bio-optical algorithm. The adoption of the proposed classification and terminology for this relatively new area for remote sensing applications is an important step for the development of this growing field.
TL;DR: In this article, a self-organizing map is used to investigate variations of the Loop Current (LC) in the Gulf of Mexico from 1992 to 2013 based on satellite-observed sea surface height data.
Abstract: The self-organizing map is used to investigate variations of the Loop Current (LC) in the Gulf of Mexico from 1992 to 2013 based on satellite-observed sea surface height data. It is found that LC variations can be characterized by three spatial patterns: normal, extension and retraction. The corresponding temporal variations confirm that LC eddy shedding generally occurs during the transition from the extension to retraction patterns. On the weekly time scale, the wind stress curl (WSC) in the Caribbean Sea has a major influence on LC eddy shedding. The increase of Caribbean WSC from June to November favours more frequent LC eddy shedding during that period. On the interannual time scale, there is also a potential linkage between the frequency of LC eddy shedding and El Nino activities.
TL;DR: Experimental results indicate that the best overall accuracy (OA) for shadow detection of the proposed object-based method was 89.60% after segmentation parameters’ optimization and scale is the most influential parameter of FNEA segmentations parameters in determining the performance of shadow detection.
Abstract: Effective treatment of shadows generated by the obstruction of trees and buildings is an inevitable task for extracting detailed spectral and spatial information from urban high-resolution images. Object-based shadow detection methods can take full advantages of spatial features in the urban very high resolution (VHR) images. However, the effect of different segmentation parameters for detecting shadows has not been well studied. In this study, we proposed an object-based method for shadow detection on urban high-resolution image and addressed quantitative assessment of segmentation. In proposed object-based method, a multi-scale segmentation method, known as fractal net evolution approach (FNEA), was employed to generate primitive objects; then, three spectral properties of shadows were fused based on Dempster–Shafer (D–S) evidence theory to identify shadows. In quantitative assessment, a method for ordering significance of parameters and deriving optimal parameters based on orthogonal experimental desig...
TL;DR: In this article, a total of 37 images of Landsat Operational Land Imager/Thematic Mapper/Enhanced Thematic Mappers plus were adopted to delineate the qualitative changes of suspended sediment concentration (SSC) in Hangzhou Bay, China.
Abstract: In this study, a total of 37 images of Landsat Operational Land Imager/Thematic Mapper/Enhanced Thematic Mapper plus were adopted to delineate the qualitative changes of Suspended Sediment Concentration (SSC) in Hangzhou Bay, China. Combined with in-situ SSC, remote sensing reflectance of the water (Rrs), water depth and simulated currents, the influence of both seabed topography and tidal currents on the SSC distribution was analysed. The results showed: (1) four High SSC Areas (HSA) and two Low SSC areas (LSA) in Hangzhou Bay. (2) SSC has a negative correlation with bathymetry, which is especially obvious during mid to late period of flood tide. HSAs appear in the shallow water (3–7 m depth) area, while the LSAs distribute in the deep water area (10–15 m depth). (3) The surface SSC distribution during the mid to late period of flood tide can help us estimate the topography information. The results of this paper can be used to other coastal embayments similar to Hangzhou Bay.
TL;DR: In this paper, the authors improved an automated land cover updating approach by integrating downscaled normalized difference vegetation index (NDVI) time series data, where the NDVI linear mixing growth model was used to detect the changed/unchanged areas and the unchanged areas.
Abstract: Land cover change monitoring is important for climate and environmental research. An automated approach for updating land cover maps derived from Landsat-like data is urgently needed to process large amounts of data. Change detection is an important part of the updating approach; however, pseudo-changes commonly occur because satellite images acquired in different seasons can capture phenological differences. Change detection based on normalized difference vegetation index (NDVI) time series data could avoid this problem; nevertheless it suffers from the much lower spatial resolution of the NDVI data. To address the resolution issue, this study improves an automated land cover updating approach by integrating downscaled NDVI time series data. First, Moderate Resolution Imaging Spectroradiometer (MODIS) NDVI data at 250-m resolution are downscaled to 30 m using the NDVI linear mixing growth model. Then, the NDVI-based change detection method is used to detect the changed/unchanged areas, and the unchanged ...
TL;DR: In this paper, the role of vegetation fraction, bedrock exposure, and slope on rocky desertification classification was analyzed using Landsat 8 Operational Land Imager (OLI) data.
Abstract: Karst rocky desertification (KRD) is a serious ecological problem in southwest China. Various remote sensing techniques are available for investigating KRD. Landsat 8 Operational Land Imager (OLI) data, acquired in September (summer) 2013 and January (winter) 2014, were used to analyse the role of vegetation fraction, bedrock exposure, and slope on KRD classification. Then, the decision tree and fuzzy maximum likelihood methods were compared, using the above-mentioned factors, to verify the potential of Landsat 8 OLI data in monitoring KRD. The results show that these factors correlate well with the degree of KRD and that the addition of these factors into the classifier improved accuracy from 84.23% to 91.71%. Thus, Landsat 8 OLI data can be adapted for the monitoring of KRD, which will be useful for the 2015 Third National Desertification Survey.
TL;DR: In order to improve the discontinuity preserving smoothing of MS filtering for cropland HRRSI, normal and uniform kernels are used to filter inner fields and boundary areas, respectively, and a new spectral bandwidth estimation is developed for better suppressing intra-field variation.
Abstract: A new image segmentation algorithm based on mean shift (MS) is proposed, with an objective to single out croplands in high-resolution remote sensing imagery (HRRSI). The algorithm is composed of two parts. First, in order to improve the discontinuity preserving smoothing of MS filtering for cropland HRRSI, normal and uniform kernels are used to filter inner fields and boundary areas, respectively. A new spectral bandwidth estimation is also developed for better suppressing intra-field variation. Second, a two-stage region-merging technique, with the second stage combining mutual best-fit rule and iterative thresholding, is implemented. An HRRSI scene is used for validation, the results of which indicate good performance of our method.
TL;DR: In this article, the authors used Lidar data for vegetation and fuel mapping in the Landscape Fire and Resource Planning Tools (LANDFIRE) program, which has become the default source of large-scale fire behaviour modelling inputs for the US.
Abstract: Accurate information about three-dimensional canopy structure and wildland fuel across the landscape is necessary for fire behaviour modelling system predictions. Remotely sensed data are invaluable for assessing these canopy characteristics over large areas; lidar data, in particular, are uniquely suited for quantifying three-dimensional canopy structure. Although lidar data are increasingly available, they have rarely been applied to wildland fuels mapping efforts, mostly due to two issues. First, the Landscape Fire and Resource Planning Tools (LANDFIRE) program, which has become the default source of large-scale fire behaviour modelling inputs for the US, does not currently incorporate lidar data into the vegetation and fuel mapping process because spatially continuous lidar data are not available at the national scale. Second, while lidar data are available for many land management units across the US, these data are underutilized for fire behaviour applications. This is partly due to a lack of local ...
TL;DR: In this article, an iteratively regularized multivariate alteration detection (IR-MAD) is applied to synthetically fused images to improve the accuracy of unsupervised change detection method and minimize registration errors among multi-temporal images.
Abstract: The main objective of this letter is to improve the accuracy of unsupervised change detection method and minimize registration errors among multi-temporal images in the change detection process. To this end, iteratively regularized multivariate alteration detection (IR-MAD) is applied to synthetically fused images. First, four synthetically fused hyperspectral images are generated using the block-based fusion method. Then, the IR-MAD is applied to three pairs of the fused images using integrated IR-MAD variates, to decrease the falsely detected changes. To focus on the mis-registration effects, we apply the method to both a correctly registered data-set and a data-set with deliberately misaligned images. In this experiment using multi-temporal Hyperion images, the changed areas are more efficiently detected by our method than by the original IR-MAD algorithm.
TL;DR: In this article, the changes in vegetation cover following the Fukushima Daiichi nuclear disaster are presented using long-term time series data obtained from Moderate Resolution Imaging Spectroradiometer (MODIS) sensors.
Abstract: The Great East Japan Earthquake and Tsunami on 11 March 2011 led to the Fukushima Daiichi nuclear disaster. The Japanese government subsequently outlined an evacuation zone around the power plant, and all residents were evacuated. In the absence of cropland or urban vegetation management, land cover was expected to change. The changes in vegetation cover following the nuclear disaster are presented using long-term time series data obtained from Moderate Resolution Imaging Spectroradiometer (MODIS) sensors. Utilizing MODIS 250 m spatial resolution observations, clear signals of vegetation changes were detected following the disaster in 2011. The areas affected were non-forest regions (mostly paddy fields) within the 20 km radius evacuation zone around the power plant. Multi-year comparisons of vegetation seasonality indicated that the changes can be explained by the natural succession of abandoned cropland. The affected area within the 20 km radius is equivalent to about 20% of the total area affected by t...
TL;DR: In this paper, a method based on a statistical analysis of the coherence distributions for wet and dry reeds using InSAR pairs was investigated, which was validated by in-situ data and showed that L-band interferometric coherence is very sensitive to the water surface conditions beneath reed marshes and so it can be used as classifier.
Abstract: Reed marshes, the world’s most widespread type of wetland vegetation, are undergoing major changes as a result of climate changes and human activities. The presence or absence of water in reed marshes has a significant impact on the whole ecosystem and remains a key indicator to identify the effective area of a wetland and help estimate the degree of degeneration. Past studies have demonstrated the use of interferometric synthetic aperture radar (InSAR) to map water-level changes for flooded reeds. However, the identification of the different hydrological states of reed marshes is often poorly understood. The analysis given in this paper shows that L-band interferometric coherence is very sensitive to the water surface conditions beneath reed marshes and so it can be used as classifier. A method based on a statistical analysis of the coherence distributions for wet and dry reeds using InSAR pairs was, therefore, investigated in this study. The experimental results were validated by in-situ data and showed...
TL;DR: The geometric hull technique proposed here performs best amongst the four feature isolation techniques and may be an important building block for next generation automatic mapping algorithms.
Abstract: Modern imaging spectrometers produce an ever-growing amount of data, which increases the need for automated analysis techniques. The algorithms employed, such as the United States Geological Survey (USGS) Tetracorder and the Mineral Identification and Characterization Algorithm (MICA), use a standardized spectral library and expert knowledge for the detection of surface cover types. Correct absorption feature definition and isolation are key to successful material identification using these algorithms. Here, a new continuum removal and feature isolation technique is presented, named the ‘Geometric Hull Technique’. It is compared to the well-established, knowledge-based Tetracorder feature database together with the adapted state of the art techniques scale-space filtering, alpha shapes and convex hull.The results show that the geometric hull technique yields the smallest deviations from the feature definitions of the MICA Group 2 library with a median difference of only 8 nm for the position of the featur...
TL;DR: In this paper, the authors used satellite-based X-band interferometric SAR (InSAR) to measure the canopy height in a tropical rainforest in Indonesia using 10 TanDEM-X InSAR data sets acquired during a 2-year period.
Abstract: Measuring canopy height using satellite-based X-band interferometric SAR (InSAR) is promising for accurate monitoring of forest biomass. A prerequisite for applying this at large scale is that the penetration of the radar microwaves into the forest canopy is stable over time, i.e. not influenced by weather conditions. We investigated this in a tropical rainforest in Indonesia using 10 TanDEM-X InSAR data sets acquired during a 2-year period. We found that mean InSAR-derived canopy height varied with a standard deviation of about 0.5 m between acquisitions. The standard variation was 0.8 m; however, about 0.3 m could be attributed to errors stemming from technical properties of the acquisitions. In conclusion, this further supports the use of X-band InSAR from satellite missions for forest monitoring.