Search

Landslides are destructive and recurrent natural disasters on steep slopes and represent a risk to lives and properties. Knowledge of relict landslides' location is vital to understand their mechanisms, update inventory maps and improve risk assessment. However, relict landslide mapping is complex in tropical regions covered with rainforest vegetation. A new CNN framework is proposed for semi-automatic detection of relict landslides, which uses a dataset generated by a k-means clustering algorithm and has a pre-training step. The weights computed in the pre-training are used to fine-tune the CNN training process. A comparison between the proposed and the standard framework is performed using CBERS-04A WPM images. Three CNNs for semantic segmentation are used (Unet, FPN and Linknet) with two augmented datasets. A total of 42 combinations of CNNs are tested. Values of precision and recall were very similar between the combinations tested. Recall was higher than 75% for every combination, but precision values were usually smaller than 20%. False positives (FP) samples were addressed as the cause for these low precision values. Predictions of the proposed framework were more accurate and correctly detected more landslides. This work demonstrates that there are limitations for detecting relict landslides in areas covered with rainforest, mainly related to similarities between the spectral response of pastures and deforested areas with Gleichenella sp. ferns, commonly used as an indicator of landslide scars. Relict landslide detection has limitations in rainforest environments The proposed framework for CNN semantic segmentation proved to be useful for relict landslide detection Land cover features spectral characteristics, and input data technical specificities are crucial to the accuracy of CNN models No optimal combination of parameters has been identified No CNN proved to be superior [ABSTRACT FROM AUTHOR]

The present study documents optimal methods for the retrieval of cloud properties using five channels (0.6, 3.7, 6.7, 10.8 and 12.0 µm) that are used in many geostationary meteorological satellite observations. Those channels are also to be adopted for the Communication, Ocean and Meteorological Satellite (COMS) scheduled to be launched in 2008. The cloud properties focused on are cloud thermodynamic phase, cloud optical thickness, effective particle radius and cloud-top properties with specific uncertainties. Discrete ordinate radiative transfer models are simulated to build up the retrieval algorithm. The cloud observations derived from the Moderate-resolution Imaging Spectroradiometer (MODIS) are compared with the results to assess the validity of the algorithm. The preliminary validation indicates that the additional use of a band at 6.7 µm would be better in discriminating the cloud ice phase. Cloud optical thickness and effective particle radius can also be produced up to, respectively, 64 and 32 µm by functionally eliminating both ground-reflected and cloud- and ground-thermal radiation components at 0.6 and 3.7 µm. Cloud-top temperature (pressure) in ±3 K (±50 hPa) uncertainties can be estimated by a simple 10.8-µm method for opaque clouds, and by an infrared ratioing method using 6.7 and 10.8 µm for semitransparent clouds. [ABSTRACT FROM AUTHOR]

The Sumatra earthquake struck South Asia on 26 December 2004 and triggered monstrous waves that turned into a tsunami hitting the ocean regions and caused the most severe natural disaster of recent decades. The devastating earthquake and tsunami changed the landscape of coastal areas in many countries around the whole Indian Ocean region. To provide real-time information for rescue and rehabilitation plans, satellite images were applied to monitor and evaluate the damage over several devastated spots. The FORMOSAT-2 satellite, which was launched on 21 May 2004 and operated by the National Space Organization, Taiwan, is uniquely designed to take timely and low-cost black and white images daily with a resolution of 2 m and colour images of 8-m resolution. FORMOSAT-2 is expected to have many useful applications, such as natural-disaster evaluation, land-usage analysis, environmental monitoring, and coastal search and rescue. FORMOSAT-2 successfully acquired several post-tsunami images of the hazardous areas, both Puhket, Thailand and Banda Aceh, Indonesia on 28 December. A series of FORMOSAT-2 satellite images were processed by geometric and radiometric correction, haze reduction, image enhancement, feature extraction, image classification, and image fusion to assess the damage over those devastated areas. FORMOSAT-2 satellite images with a high-temporal resolution and high-spatial resolution were proved to be an efficient and useful information source for decision-makers to make rescue and recovery plans, especially for some isolated islands hard to reach in time. [ABSTRACT FROM AUTHOR]

Optically thin cirrus derived from the Moderate‐resolution Imaging Spectroradiometer (MODIS) instrument is examined in connection with its cloud phase (ice crystal or liquid droplet) for quasi‐single layer thin cirrus (either single‐layer thin cirrus or thin cirrus overlying low‐level thin clouds) over water surface conditions. Analysing 142 MODIS data samples during the period 1–16 March 2000, it is found that only about 20% of MODIS thin cirrus having a total‐column cloud optical depth less than 2.0 was detected as ice phase. In the present study, MODIS cloud phase datasets are newly reproduced by adding the 1.6 µm reflectance to an existing infrared trispectral algorithm of MODIS which used the 8.7, 11 and 12 µm bands. The percentage of ice phase in the thin cirrus is estimated at about 80% when new cloud phase is incorporated. The increase is found for all regions, in particular in mid‐latitudes. [ABSTRACT FROM AUTHOR]

This article focuses on the Serra da Cangalha astrobleme. Structural features caused by the impact of celestial bodies at the surface of the Earth always stir scientific interest. These impacts instantaneously release huge quantities of kinetic energy that depends only on the mass and the velocity of the cosmic body. The Serra da Cangalha astrobleme is the best preserved impact crater found in Brazil, as depicted by two modern sources of orbital remote sensing data: multispectral images acquired by the Advanced Spaceborne Thermal Emission and Reflection Radiometer, and terrain elevation data collected by the Shuttle Radar Topographic Mission.

This paper investigates the potential for utilizing multi-frequency, quad-polarized Synthetic Aperture Radar (SAR) for delineation of fire scars in the tropical savannas of northern Australia. Fire regimes and regional land management in the study area are contentious issues for the multiple stakeholders. Thus, the ability to accurately map fire scars and assess impacts on the mosaic of vegetation community types is crucial for effective landscape management. In this study identification of a fire scar resulting from a late dry season fire was assessed using TopSAR data. The assessment was achieved through comparison with optical datasets acquired both pre-burn and post-burn. By examining five representative vegetation communities, it was concluded that only C-band SAR data were affected sufficiently by fire in this environment to detect and map fire scars. Despite the intensity of the fire event, the resistance of vegetation communities to fire damage resulted in insignificant changes to the L- and P-band SAR data. Further investigation is required to determine if this behaviour can be exploited to improve the above-ground woody biomass mapping currently reliant on optical data. [ABSTRACT FROM AUTHOR]

The variability of sea surface temperature in the region of the Kuroshio intrusion into the South China Sea (SCS) through the Luzon Strait was studied using sea surface temperature (SST) derived from Advanced Very High Resolution Radiometer (AVHRR) from 1985 to 2002. The covariance empirical orthogonal function (CEOF) method was applied for analysing the temporal and spatial variability in the study area. The results show that the Kuroshio intrusion during El Niño periods is weaker than that in La Niña periods. The calculation of surface layer heat of the Kuroshio intrusion region also shows response to the El Niño-La Niña events. The variation is attributed to the changes in wind fields during those events. [ABSTRACT FROM AUTHOR]

The selection of features, including spectral, texture, shape, size, and signal strength, is an important step in computerized information analysis of remotely sensed images. A feature space, which can be generally understood as a multidimensional space consisting of multiple individual features, can be modelled by estimating the distribution of the whole space with prior assumed probability distribution functions (PDFs) once only. However, due to the inter-overlapping phenomenon among points or the confusing influence from surrounding discrete points, it is very difficult to obtain the subtle and procedural structure of the mixture distributions of feature space, and so as to degrade the accuracy and interpretability of the results in further analysis. Extending on the method of Gaussian mixture modelling and decomposition (GMDD), a new feature mining method--stepwise optimization model (SOM) with genetic algorithms (GA) was proposed in this study for the extraction of tree-like hierarchical structure of unknown feature distributions in a feature space. To approximate reality accurately, integration of SOM-GA with symbolic geographical knowledge is essential in the feature mining and classification of remotely sensed images. Knowledge-integrated SOM-GA model that combines the power of SOM-GA and logic reasoning of rule-based inference was therefore proposed. The paper presents conceptual and technical discussions of the model in detail, along with the result of practical application test on a district in Hong Kong region. [ABSTRACT FROM AUTHOR]

A detailed assessment of the planimetric accuracy of high-resolution panchromatic Ikonos orthoimage products for three different test sites is presented. The main objective was to evaluate the potential of Ikonos orthoimage products for use as digital image basemaps in local government Geographic Information Systems (GIS) systems. For maximum utility and adoption, a planimetric accuracy of 3-5 m and a circular error at 90% probability (CE90) is considered nominal for local government planning and management applications. Low-precision, low-cost georeferenced Ikonos image products (Geo, 50 m CE90) were orthorectified using third party commercially available software and various custom Digital Elevation Models (DEMs). The planimetric accuracies of the resulting custom Ikonos orthoimages were found to vary between 2-4 m CE90. In addition, United States Geological Survey (USGS) DEMs were also used to orthorectify georeferenced Ikonos image products. Planimetric accuracies of 2-3 m CE90 were obtained from georeferenced Ikonos using USGS DEMs with RMS vertical accuracies of the order of 2 m. The approach demonstrated here can be used to deliver up-to-date, cost-effective orthoimages from Ikonos Geo products that yield planimetric accuracies suitable for use as digital image basemaps by local governments. [ABSTRACT FROM AUTHOR]

The aim of this study was to find the most appropriate method of classification for the Thames intertidal habitat types at Crayford Marsh and Dartford Creek by using Compact Airborne Spectrographic Imager (CASI ) data. Preliminary evaluation of commonly available classification algorithms produced two candidate techniques: the Maximum Likelihood Classifier (MLC) and the Spectral Angle Mapper (SAM). Pre-classification enhancements and the two different classifiers were compared. Ten different dataset combinations were created for two pilot sites: one at Crayford Marsh and one at Dartford Creek. These consisted of the original CASI bandset (15 bands in spatial mode from blue to near-infrared) and nine other combinations resulting from band subsets, Principal Component Analysis (PCA) and Normalized Difference Vegetation Index (NDVI ). Twelve classes were established for each site although only some of these were common to both. Each classified image was accuracy assessed using a combination of field mapping, field photographs and air photograph interpretation as reference data. The most accurate classification (68% for Crayford Marsh and 53% for Dartford Creek) for both sites comprised the use of MLC with a dataset created from PCs 2, 3 and 4 from a PCA carried out on the original 15 band data, combined with an additional NDVI band. CASI data proved useful for the mapping of salt-marsh vegetation and sediments especially in the Crayford Marsh site. In the Dartford Creek site, however, there was significant confusion between some classes. Further work is recommended to test the classification. [ABSTRACT FROM AUTHOR]

Recognition of objects extracted from remotely sensed imagery requires the matching of object properties with prior stored knowledge. Various properties were used in this study to form the model of a priori knowledge. A GIS (geographical information system) dataset was used to assist the extraction of shape descriptors, reflectance and height above ground characteristics of classes of object including building, road, grassland and tree. Human interpreters are capable of recognizing objects in natural scenes (including aerial photography) that display complex, overlapping composition and representation. Objects extracted from such imagery are inherently fuzzy. In order to perform the recognition task by computer, such uncertainty must be accommodated. Many researchers have used the robustness of neural networks to accomplish such recognition. In this work, we utilized a fuzzy expert system and an adaptive neuro-fuzzy system to train, adapt and recognize objects in three complex aerial scenes. [ABSTRACT FROM AUTHOR]

Radar imaging is a valuable tool for the monitoring and management of tropical ecosystems. One of the obstacles to its successful usage is the variation of the radar backscatter with incidence angle caused by the side-looking mode of the sensor. Available research has modelled the effect of changes in incidence angle for soil, grass and scattering mechanisms in tree-covered areas. The effect varies depending on the scattering mechanism. To model the radar backscatter behaviour of a natural landscape, therefore, implies precise knowledge of the land-cover composition. A method for estimating and correcting the effect of changes in look angle on backscatter data is consequently proposed; it requires little field knowledge and encompasses the effects on most land-cover types in the study area. The method is based on the assumption that each line in azimuth direction contains a similar composition in regard to land-cover types. The backscatter frequency distribution of each azimuth line can then be employed to model and correct for the effect of variation in incidence angle. The results are evaluated using an existing land-cover classification to extract mean backscatter values for individual land-cover classes before and after the correction procedure. The correction procedure is shown to successfully adjust backscatter intensities to a nominal incidence angle for the vegetation communities in a coastal tropical savanna landscape in Australia's Northern Territory. [ABSTRACT FROM AUTHOR]

Very high resolution airborne video data were used to assess the optimal spatial resolution for mapping the land cover types occurring in the tropical savannas of northern Australia. Canopy cover and principal component images were extracted for four vegetation communities and investigated separately. Analysis of average variance in conjunction with a proposed visual analysis method proved more successful in establishing the optimal resolution than analysis of spatial autocorrelation using semivariograms. The optimal resolution was found to be dependent on the structure of the vegetation community as well as the image component under investigation. A resolution of between 20 and 27 m is optimal for all instances examined. [ABSTRACT FROM AUTHOR]

In this paper we present an empirical relationship between the broadband glacier albedo (alpha) and the narrowband glacier albedos in Landsat TM bands 2 and 4 (alpha2 and alpha4, respectively). The relationship was established on the basis of multiple linear regression analysis of 112 ground-based simultaneous measurements of alpha, alpha2 and alpha4 made at 32 sites on the tongue of the Morteratschgletscher, Switzerland. The measurements were carried out over a representative set of glacier surface types ranging from completely debris-covered glacier ice (alpha=0.08)to dry snow (alpha=0.86). The regression model explains more than 99% of the variance of the broadband albedo and the root-mean-square value of the residuals is only 0.009. The relationship enables users of Landsat TM data to make an accurate estimate of the broadband albedo on the basis of narrowband albedos without having to classify the glacier surface. [ABSTRACT FROM AUTHOR]

Abstract. Inshore reefs located on the coastline of northern Australia are foraging grounds for several threatened species of sea turtle which feed on brown algae present on the reef flat. Using Landsat TM visible bands, an iterative digital image classification is used to delineate the major reef flat zones. In addition, TM band 4 is used to assist in the identification of algal growth, showing that this band can successfully separate substrate types up to a depth of 35cm in clear calm water. The classification results reveal that the TM data have the potential to identify crucial turtle feeding grounds along the vast coastline of northern Australia. [ABSTRACT FROM AUTHOR]