Land-cover classification in a moist tropical region of Brazil with Landsat Thematic Mapper imagery

This research aims to improve land-cover classification accuracy in a moist tropical region in Brazil by examining the use of different remote-sensing-derived variables and classification algorithms. Different scenarios based on Landsat Thematic Mapper (TM) spectral data and derived vegetation indices and textural images and different classification algorithms, maximum likelihood classification (MLC), artificial neural network (ANN), classification tree analysis (CTA) and object-based classification (OBC), were explored. The results indicate that a combination of vegetation indices as extra bands into Landsat TM multi-spectral bands did not improve the overall classification performance, but the combination of textural images was valuable for improving vegetation classification accuracy. In particular, the combination of both vegetation indices and textural images into TM multi-spectral bands improved the overall classification accuracy (OCA) by 5.6% and the overall kappa coefficient (OKC) by 6.25%. Comparison of the different classification algorithms indicated that CTA and ANN have poor classification performance in this research, but OBC improved primary forest and pasture classification accuracies. This research indicates that use of textural images or use of OBC are especially valuable for improving the vegetation classes such as upland and liana forest classes that have complex stand structures and large patch sizes.     

Land-cover classification in a moist tropical region of Brazil with Landsat TM imagery

This research aims to improve land-cover classification accuracy in a moist tropical region in Brazil by examining the use of different remote sensing-derived variables and classification algorithms. Different scenarios based on Landsat Thematic Mapper (TM) spectral data and derived vegetation indices and textural images, and different classification algorithms - maximum likelihood classification (MLC), artificial neural network (ANN), classification tree analysis (CTA), and object-based classification (OBC), were explored. The results indicated that a combination of vegetation indices as extra bands into Landsat TM multispectral bands did not improve the overall classification performance, but the combination of textural images was valuable for improving vegetation classification accuracy. In particular, the combination of both vegetation indices and textural images into TM multispectral bands improved overall classification accuracy by 5.6% and kappa coefficient by 6.25%. Comparison of the different classification algorithms indicated that CTA and ANN have poor classification performance in this research, but OBC improved primary forest and pasture classification accuracies. This research indicates that use of textural images or use of OBC are especially valuable for improving the vegetation classes such as upland and liana forest classes having complex stand structures and having relatively large patch sizes.     

Using multiscale texture information from ALOS PALSAR to map tropical forest

This research investigated the ability of the Advanced Land Observing Satellite (ALOS) Phased Array type L-band Synthetic Aperture Radar (PALSAR) to map tropical forest in central Sumatra, Indonesia. The study used PALSAR 50 m resolution orthorectified HH and HV data. As land-cover discrimination is difficult with only two bands (HH and HV), we added textures as additional information for classification. We calculated both first- and second-order texture features and studied the effects of texture window size, quantization scale and displacement length on discrimination capability. We found that rescaling to a lower number of grey levels (8 or 16) improved discrimination capability and that equal probability quantization was more effective than uniform quantization. Increasing displacement tended to reduce the discrimination capability. Low spatial resolution increased the discrimination capability because low spatial resolution features reduce the effects of noise. A larger number of features also improved discrimination capability. However, the amount of improvement depended on the window size. We used the optimum combination of backscatter amplitude and textures as input data into a supervised multi-resolution maximum likelihood classification. We found that including texture information improved the overall classification accuracy by 10%. However, there was significant confusion between natural forest and acacia plantations, as well as between oil palm and clear cuts, presumably because the backscatter and texture of these class pairs are very similar.     

Land-cover mapping in the Nujiang Grand Canyon: integrating spectral,textural, and topographic data in a random forest classifier

The integration of spectral, textural, and topographic information using a random forest classifier for land-cover mapping in the rugged Nujiang Grand Canyon was investigated in this study. Only a few land-cover categories were accurately discriminated using spectral information exclusively, with an overall accuracy of 0.56 and a kappa coefficient of 0.51. The inclusion of topographic information as additional bands provided higher overall accuracy (0.69) and kappa coefficient (0.65) than topographic correction (overall accuracy, 0.57–0.58; kappa coefficient range, 0.52–0.53), which failed to markedly improve classification accuracy. In contrast with the exclusive use of spectral bands, most of the included land-cover categories were correctly classified using textural features exclusively (overall accuracy, 0.67–0.88; kappa coefficient, 0.63–0.87). In particular, classification based on geostatistical features led to slightly more accurate results than did grey-level co-occurrence matrix parameters. The window size selected for texture calculation markedly affected the texture-based classification accuracy: larger window size yielded higher classification accuracy. However, no optimal window size exists. The inclusion of the topographic bands in the texture images led to an increase in the overall accuracy of 1.1–9.0%, and to an increase in the kappa coefficient of 0.0–10.9%. Thus, for the Nujiang Grand Canyon, topographic information was more important for the discrimination of some land-cover types than spectral and textural information. Among the Landsat Thematic Mapper (TM) spectral bands, bands 6 and 4 were of greatest importance. The relative importance of textural features generally increased with window size, and a few textural features were of consistently high importance. Although a random forest classifier does not overfit, undertaking feature selection analysis prior to classification may still be valuable.     

Towards a global oil palm sample database: design and implications

Global oil palm plantations have expanded in the last few decades, resulting in negative impacts on the environment. Satellite remote sensing plays an important role in monitoring the expansion of oil palm plantations, but requires high-quality ground samples for training and validation. To facilitate the monitoring of oil palm plantations on a large scale, we propose an oil palm sample database that includes the five countries with the largest areas of oil palm plantations: Indonesia, Malaysia, Nigeria, Thailand, and Ghana. In total, 45,896 samples were collected using a hexagonal sampling design. High-resolution images from Google Earth, the Advanced Land Observing Satellite (ALOS) Phased Array type L-band Synthetic Aperture Radar (PALSAR) images, and Landsat optical images were used to identify oil palm plantations and other types of land cover (croplands, forests, grasslands, shrublands, water, hard surfaces, and bare land). The characteristics of oil palm cover and its environment, including PALSAR backscattering coefficients, terrain, and climate recorded in this database are also discussed. The results indicate that using the PALSAR band algebra threshold alone is not recommended to distinguish oil palm from other land-cover/use types.     

Mapping the land-cover distribution in arid and semiarid urban landscapes with Landsat Thematic Mapper imagery

Mapping the land-cover distribution in arid and semiarid urban landscapes using medium spatial resolution imagery is especially difficult due to the mixed-pixel problem in remotely sensed data and the confusion of spectral signatures among bare soils, sparse density shrub lands, and impervious surface areas (ISAs hereafter). This article explores a hybrid method consisting of linear spectral mixture analysis (LSMA), decision tree classifier, and cluster analysis for mapping land-cover distribution in two arid and semiarid urban landscapes: Urumqi, China, and Phoenix, USA. The Landsat Thematic Mapper (TM) imagery was unmixed into four endmember fraction images (i.e. high-albedo object, low-albedo object, green vegetation (GV), and soil) using the LSMA approach. New variables from these fraction images and TM spectral bands were used to map seven land-cover classes (i.e. forest, shrub, grass, crop, bare soil, ISA, and water) using the decision tree classifier. The cluster analysis was further used to modify the classification results. QuickBird imagery in Urumqi and aerial photographs in Phoenix were used to assess classification accuracy. Overall classification accuracies of 86.0% for Urumqi and 88.7% for Phoenix were obtained, much higher accuracies than those utilizing the traditional maximum likelihood classifier (MLC). This research demonstrates the necessity of using new variables from fraction images to distinguish between ISA and bare soils and between shrub and other vegetation types. It also indicates the different effects of spatial patterns of land-cover composition in arid and semiarid landscapes on urban land-cover classification.     

Using MERIS fused images for land-cover mapping and vegetation status assessment in heterogeneous landscapes

In this paper we evaluate the potential of ENVISAT–Medium Resolution Imaging Spectrometer (MERIS) fused images for land-cover mapping and vegetation status assessment in heterogeneous landscapes. A series of MERIS fused images (15 spectral bands; 25 m pixel size) is created using the linear mixing model and a Landsat Thematic Mapper (TM) image acquired over the Netherlands. First, the fused images are classified to produce a map of the eight main land-cover types of the Netherlands. Subsequently, the maps are validated using the Dutch land-cover/land-use database as a reference. Then, the fused image with the highest overall classification accuracy is selected as the best fused image. Finally, the best fused image is used to compute three vegetation indices: the normalized difference vegetation index (NDVI) and two indices specifically designed to monitor vegetation status using MERIS data: the MERIS terrestrial chlorophyll index (MTCI) and the MERIS global vegetation index (MGVI).

Results indicate that the selected data fusion approach is able to downscale MERIS data to a Landsat-like spatial resolution. The spectral information in the fused images originates fully from MERIS and is not influenced by the TM data. Classification results for the TM and for the best fused image are similar and, when comparing spectrally similar images (i.e. TM with no short-wave infrared bands), the results of the fused image outperform those of TM. With respect to the vegetation indices, a good correlation was found between the NDVI computed from TM and from the best fused image (in spite of the spectral differences between these two sensors). In addition, results show the potential of using MERIS vegetation indices computed from fused images to monitor individual fields. This is not possible using the original MERIS full resolution image. Therefore, we conclude that MERIS–TM fused images are very useful to map heterogeneous landscapes.     

An evaluation of Landsat TM spectral data and SAR-derived textural information for lithological discrimination in the Red Sea Hills,Sudan

The effectiveness of spectral and textural information in the identification of surface rock types in an arid region, the Red Sea Hills of Sudan, is evaluated using spectral information from the six Landsat TM optical bands and textural features derived from Shuttle Imaging Radar-C (SIR-C) C-band HH polarization data. An initial classification is derived from Landsat TM data alone using three classification algorithms, Gaussian maximum likelihood, a multi-layer feed-forward neural network and a Kohonen self-organizing feature map (SOM), to generate lithological maps, with classification accuracy being measured using a confusion matrix approach. The feed-forward neural net produced the highest overall classification accuracy of 57 per cent and was, therefore, selected for the second experiment, in which texture measures from SIR-C C-band HH-polarized synthetic aperture radar (SAR) data are added to selected TM spectral features. Four methods of measuring texture are employed, based on the Fourier power spectrum, grey level co-occurrence matrix (GLCM), multi-fractal measures, and the multiplicative autoregressive random field (MAR) model. The use of textural information together with a subset of the TM spectral features leads to an increase in classification accuracy to almost 70 per cent. Both the MAR model and the GLCM matrix approach perform better than Fourier and multi-fractal based methods of texture characterization.     

Obtaining crop-specific time profiles of NDVI: the use of unmixing approaches for serving the continuity between SPOT-VGT and PROBA-V time series

The study examined the potential of two unmixing approaches for deriving crop-specific normalized difference vegetation index (NDVI) profiles so that upon availability of Project for On-Board Autonomy – Vegetation (PROBA-V) imagery in winter 2013, this new data set can be combined with existing Satellite Pour l’Observation de la Terre – VEGETATION (SPOT-VGT) data despite the differences in spatial resolution (300 m of PROBA-V versus 1 km of SPOT-VGT). To study the problem, two data sets were analysed: (1) a set of 10 temporal NDVI images, with 300 and 1000 m spatial resolution, from the state of São Paulo (Brazil) synthesized from 30 m Landsat Thematic Mapper (TM) images, and (2) a corresponding set of 10 observed Moderate Resolution Imaging Spectroradiometer (MODIS) images (250 m spatial resolution). To mimic the influence of noise on the retrieval accuracy, different sensor/atmospheric noise levels were applied to the first data set. For the unmixing analysis, a high-resolution land-cover (LC) map was used. The LC map was derived beforehand using a different set of Landsat TM images. The map distinguishes nine classes, with four different sugarcane stages, two agricultural sub-classes, plus forest, pasture, and urban/water. Unmixing aiming at the retrieval of crop-specific NDVI profiles was done at administrative level. For the synthesized data set it was demonstrated that the ‘true’ NDVI temporal profiles of different land-cover classes (from 30 m TM data) can generally be retrieved with high accuracy. The two simulated sensors (PROBA-V and SPOT-VGT) and the two unmixing algorithms gave similar results. Analysing the MODIS data set, we also found a good correspondence between the modelled NDVI profiles (both approaches) and the (true) Landsat temporal endmembers.     

Mapping oil palm extent in Malaysia using ALOS-2 PALSAR-2 data

The extent of oil palm plantations has increased rapidly in Malaysia over the past few decades. To evaluate ecological effects and economic values, it is important to produce an accurate oil palm map for Malaysia. The Phased Array Type L-band Synthetic Aperture Radar (PALSAR) on the Advance Land Observing Satellite (ALOS) is useful in land-cover mapping in tropical regions under all-weather conditions. In this study, PALSAR-2 images from 2015 were used for oil palm mapping with maximum likelihood classifier (MLC)-based supervised classification. The processed PALSAR-2 data were resampled to multiple coarser resolutions (50, 100, 250, 500, and 1000 m), and then used to investigate the effect of speckle in oil palm mapping. Both independent testing samples and inventories from the Malaysia Palm Oil Board (MPOB) were used to evaluate the mapping accuracy. The oil palm mapping result indicates 50–500 m to be a good resolution for either retaining spatial details or reducing speckle noise of PALSAR-2 images. Among which, the best overall mapping accuracies and average oil palm accuracies reached 94.50% and 89.78%, respectively. Moreover, the oil palm area derived from the 100-m resolution map is 6.14 million hectares (Mha), which is the closest to the official MPOB inventories (~8.87% overestimation).     

Use of UK-DMC 2 and ALOS PALSAR for studying the age of oil palm trees in southern peninsular Malaysia

This article demonstrates some techniques for studying the age of oil palm trees (Elaeis guineensis Jacq.) using the Disaster Monitoring Constellation 2 from the UK (UK-DMC 2) and Advanced Land Observing Satellite phased array L-band synthetic aperture radar (ALOS PALSAR) remote-sensing data at a private oil palm estate in southern peninsular Malaysia. Several techniques were explored with UK-DMC 2 data, namely (1) radiance, vegetation indices, and fraction of shadow; (2) texture measurement; (3) classifications, namely Iterative Self-Organizing Data Analysis Technique (ISODATA) classification, maximum-likelihood classification (MLC), and random forest (RF) classification; (4) in terms of ALOS PALSAR data, the correlation of polarizations (i.e. horizontal transmitting and horizontal receiving (termed HH polarization) and horizontal transmitting and vertical receiving (termed HV polarization)) and the ratio of these polarizations to the age of oil palm trees. From the results, band 1 (near-infrared) of UK-DMC 2, fraction of shadow, and mean filter from the grey-level co-occurrence matrix (GLCM) demonstrated strong correlation of determination (R ²?=?0.76–0.80) with the age of oil palm trees, while the ALOS PALSAR HH polarization could correlate moderately strongly (R ²?=?0.49) with the age of oil palm trees. Adding fraction of shadow and UK-DMC 2 data using the RF method further improved the overall accuracy of age classification from 45.3% (MLC method) to 52.9%. This study concluded that texture measurement (GLCM mean) and fraction of shadow are useful for studying the age of oil palm trees, although discriminating variation in age between mature oil palm trees is difficult because the leaf area index development of mature oil palm trees stabilizes at about 10 years of age. Future studies should involve height information, because this has the potential to be used as one of the most important variables for studying the age of oil palm trees.     

Mapping paddy rice with multitemporal ALOS/PALSAR imagery in southeast China

Mapping rice cropping areas with optical remote sensing is often a challenge in tropical and subtropical regions because of frequent cloud cover and rainfall during the rice growing season. Synthetic aperture radar (SAR) is a potential alternative for rice mapping because of its all-weather imaging capabilities. The recent Phased Array-type L-band SAR (PALSAR) sensor onboard the Advanced Land Observing Satellite (ALOS) acquires multipolarization and multitemporal images that are highly suitable for rice mapping. In this pilot study, we demonstrate the feasibility of this sensor in mapping the rice planting area in Zhejiang Province, southeast China. High-resolution ALOS/PALSAR images were acquired at three rice growing stages (transplanting, tillering and heading) and were applied in a support vector machine (SVM) classifier to map rice and other land use surfaces. The results show that, based on the 1:10 000 land use/land cover (LULC) survey map, the rice fields can be mapped with a conditional Kappa value of 0.87 and at user's and producer's accuracies of 90% and 76%, respectively. The large commission error primarily came from confusion between rice and dryland crops or orchards because of their similar backscatter amplitudes in the rice growing season. The relatively high rice mapping accuracy in this study indicates that the new ALOS/PALSAR data could provide useful information in rice cropping management in subtropical regions such as southeast China.     

Automatic mapping of soil texture through the integration of ground,satellite and ancillary data

Satellite images obtained in the optical domain can provide information on important soil properties, such as texture. The use of these images to automatically map soil texture is, however, complicated by the presence of vegetation cover, which can mask the soil spectral response. A multistep methodology based on the use of ground, satellite and ancillary data is proposed and tested to map soil texture in Grosseto, a province of Central Italy. The methodology first separated vegetated and nonvegetated pixels of Landsat Thematic Mapper (TM) images by the use of an appropriate spectral index, the Soil Adjusted Vegetation Index (SAVI). Next, different transforms (nonparametric and parametric) were tuned using ground samples and applied to the two pixel types to separately extract relevant spectral information. The outcomes of these transforms were then merged and subjected to further processing aimed at reducing noise and conveying spatial information to the mapping process. The stratification of the soil texture estimates obtained on different lithological units was finally tested to further improve map accuracy.     

Annual land-cover mapping based on multi-temporal cloud-contaminated landsat images

Landsat images, which have fine spatial resolution, are an important data source for land-cover mapping. Multi-temporal Landsat classification has become popular because of the abundance of free-access Landsat images that are available. However, cloud cover is inevitable due to the relatively low temporal frequency of the data. In this paper, a novel approach for multi-temporal Landsat land-cover classification is proposed. The land cover for each Landsat acquisition date was first classified using a Support Vector Machine (SVM) and then the classification results were combined using different strategies, with missing observations allowed. Three strategies, including the majority vote (MultiSVM-MV), Expectation Maximisation (MultiSVM-EM) and joint SVM probability (JSVM), were used to merge the multi-temporal classification maps. The three algorithms were then applied to a region of the path/row 143/31 scene using 2010 Landsat-5 Thematic Mapper (TM) images. The results demonstrated that, for these three algorithms, the average overall accuracy (OA) improved with the increase in temporal depth; also, for a given temporal depth, the performance of JSVM was clearly better than that of MultiSVM-MV and MultiSVM-EM, and the performance of MultiSVM-EM was slightly better than that of MultiSVM-MV. The OA values for the three classification results, which use all epochs, were 70.28%, 72.40% and 74.80% for MultiSVM-MV, MultiSVM-EM and JSVM, respectively. In comparison, two other annual composite image-based classification methods, annual maximum Normalised Difference Vegetation Index (NDVI) composite image-based classification and annual best-available-pixel (BAP) composite image-based classification, gave OA values of 68.08% and 69.92%, respectively, meaning that our method produced a better performance. Therefore, the novel multi-temporal Landsat classification method presented in this paper can deal with the cloud-contamination problem and produce accurate annual land-cover mapping using multi-temporal cloud-contaminated images, which is of importance for regional and global land-cover mapping.     

Fraction images in multitemporal change detection

The concept of mixed pixels allows the interpretation of remote sensing digital image data at sub-pixel level. Fraction-image data, obtained using the notion of mixed pixels, offer a potentially powerful method to detect changes in land-cover over a given period of time. This study proposes a new approach to detect land-cover changes, using two sets of fraction-image data obtained from sets of multispectral image data acquired at two different dates, over the same area. Changes based on the selected pixel components are then used to generate the fraction-change image data, including both positive (increase) and negative (decrease) changes in each component. The proposed analysis is then performed in the fraction-change space in two different ways: (1) by implementing unsupervised classification methods and (2) by comparing the fraction-change images among themselves. The proposed methodology is tested on two sets of Landsat Thematic Mapper (TM) multispectral image data obtained at two different dates and covering a test area mapped in previous works. Results obtained by the proposed methodology are presented and discussed.     

Land cover estimation with ALOS satellite image using a neural-network

On May 12, 2008, a large earthquake occurred in Sichuan, China. We analyzed the damage caused by this disaster using satellite images from ALOS, a Japanese satellite. The land cover classification is operated by images captured on AVNIR-2. Frequently, the AVNIR-2 images cannot be monitored because of the cloud cover and solar irradiation. The area near the center of the earthquake area is covered with clouds. The goal of this article is to classify the land cover using PALSAR images. PALSAR can observe over a 350-km-wide area independently of the weather. The PALSAR is a single-band sensor, and the inputs consist of many pixels by using the nearest pixel values, and the supervisor signal is the classes estimated by AVNIR-2.     

ALOS/PALSAR InSAR数学模型研究

日本ALOS卫星携带的相控阵型L波段合成孔径雷达(PALSAR),因其较长的波长使得相同时间间隔内地面具有较高的相干性,因而极具InSAR应用潜力。本文首先介绍ALOS PALSAR,进而详细分析该数据在InSAR数学模型(包括距离向频谱、干涉临界基线距、模糊高度、差分相位对形变的敏感度)中的特点,并与常见的ERS SAR数据进行比较。     

A maximum entropy method to extract urban land by combining MODIS reflectance,MODIS NDVI,and DMSP-OLS data

Researchers often encounter difficulties in obtaining timely and detailed information on urban growth. Modern remote-sensing techniques can address such difficulties. With desirable spectral resolution and temporal resolution, Moderate Resolution Imaging Spectroradiometer (MODIS) products have significant advantages in tackling land-use and land-cover change issues at regional and global scales. However, simply based on spectral information, traditional methods of remote-sensing image classification are barely satisfactory. For example, it is quite difficult to distinguish urban and bare lands. Moreover, training samples of all land-cover types are needed, which means that traditional classification methods are inefficient in one-class classification. Even support vector machine, a current state-of-the-art method, still has several drawbacks. To address the aforementioned problems, this study proposes extracting urban land by combining MODIS surface reflectance, MODIS normalized difference vegetation index (NDVI), and Defense Meteorological Satellite Program Operational Linescan System data based on the maximum entropy model (MAXENT). This model has been proved successful in solving one-class problems in many other fields. But the application of MAXENT in remote sensing remains rare. A combination of NDVI and Defense Meteorological Satellite Program Operational Linescan System data can provide more information to facilitate the one-class classification of MODIS images. A multi-temporal case study of China in 2000, 2005, and 2010 shows that this novel method performs effectively. Several validations demonstrate that the urban land extraction results are comparable to classified Landsat TM (Thematic Mapper) images. These results are also more reliable than those of MODIS land-cover type product (MCD12Q1). Thus, this study presents an innovative and practical method to extract urban land at large scale using multiple source data, which can be further applied to other periods and regions.     

Mapping and monitoring riparian vegetation distribution,structure and composition with regression tree models and post-classification change metrics

Riparian systems have become increasingly susceptible to both natural and human disturbances as cumulative pressures from changing land use and climate alter the hydrological regimes. This article introduces a landscape dynamics monitoring protocol that incorporates riparian structural classes into the land-cover classification scheme and examines riparian change within the context of surrounding land-cover change. We tested whether Landsat Thematic Mapper (TM) imagery could be used to document a riparian tree die-off through the classification of multi-date Landsat images using classification and regression tree (CART) models trained with physiognomic vegetation data. We developed a post-classification change map and used patch metrics to examine the magnitude and trajectories of riparian class change relative to mapped disturbance parameters. Results show that catchments where riparian change occurred can be identified from land-cover change maps; however, the main change resulting from the die-off disturbance was compositional rather than structural, making accurate post-classification change detection difficult.