Satellite Image Information Research Articles

Data-Driven Earthquake Multi-impact Modeling: A Comparison of Models

In this study, a broad range of supervised machine learning and parametric statistical, geospatial, and non-geospatial models were applied to model both aggregated observed impact estimate data and satellite image-derived geolocated building damage data for earthquakes, via regression- and classification-based models, respectively. For the aggregated observational data, models were ranked via predictive performance of mortality, population displacement, building damage, and building destruction for 375 observations across 161 earthquakes in 61 countries. For the satellite image-derived data, models were ranked via classification performance (damaged/unaffected) of 369,813 geolocated buildings for 26 earthquakes in 15 countries. Grouped k-fold, 3-repeat cross validation was used to ensure out-of-sample predictive performance. Feature importance of several variables used as proxies for vulnerability to disasters indicates covariate utility. The 2023 Türkiye–Syria earthquake event was used to explore model limitations for extreme events. However, applying the AdaBoost model on the 27,032 held-out buildings of the 2023 Türkiye–Syria earthquake event, predictions had an AUC of 0.93. Therefore, without any geospatial, building-specific, or direct satellite image information, this model accurately classified building damage, with significantly improved performance over satellite image trained models found in the literature.

Improving landslide prediction by computer vision and deep learning

The destructive power of a landslide can seriously affect human beings and infrastructures. The prediction of this phenomenon is of great interest; however, it is a complex task in which traditional methods have limitations. In recent years, Artificial Intelligence has emerged as a successful alternative in the geological field. Most of the related works use classical machine learning algorithms to correlate the variables of the phenomenon and its occurrence. This requires large quantitative landslide datasets, collected and labeled manually, which is costly in terms of time and effort. In this work, we create an image dataset using an official landslide inventory, which we verified and updated based on journalistic information and interpretation of satellite images of the study area. The images cover the landslide crowns and the actual triggering values of the conditioning factors at the detail level (5 × 5 pixels). Our approach focuses on the specific location where the landslide starts and its proximity, unlike other works that consider the entire landslide area as the occurrence of the phenomenon. These images correspond to geological, geomorphological, hydrological and anthropological variables, which are stacked in a similar way to the channels of a conventional image to feed and train a convolutional neural network. Therefore, we improve the quality of the data and the representation of the phenomenon to obtain a more robust, reliable and accurate prediction model. The results indicate an average accuracy of 97.48%, which allows the generation of a landslide susceptibility map on the Aloag-Santo Domingo highway in Ecuador. This tool is useful for risk prevention and management in this area where small, medium and large landslides occur frequently.

Analysis of Spatial–Temporal Variability of PM2.5 Concentrations Using Optical Satellite Images and Geographic Information System

Particulate matter less than 2.5 microns in diameter (PM2.5) is an air pollutant that has become a major environmental concern for governments around the world. Management and control require air quality monitoring and prediction. However, previous studies did not fully utilize the spectral information in multispectral satellite images and land use data in geographic datasets. To alleviate these problems, this study proposes the extraction of land use information not only from geographic inventory but also from satellite images with a machine learning-based classification. In this manner, near up-to-date land use data and spectral information from satellite images can be utilized, and the integration of geographic and remote sensing datasets boosts the accuracy of PM2.5 concentration modeling. In the experiments, Landsat-8 imagery with a 30-m spatial resolution was used, and cloud-free image generation was performed prior to the land cover classification. The proposed method, which uses predictors from geographic and multispectral satellite datasets in modeling, was compared with an approach which utilizes geographic and remote sensing datasets, respectively. Quantitative assessments showed that the proposed method and the developed model, with a performance of RMSE = 3.06 µg/m3 and R2 = 0.85 comparatively outperform the models with a performance of RMSE = 3.14 µg/m3 and R2 = 0.68 for remote sensing datasets and a performance of RMSE = 3.47 µg/m3 and R2 = 0.79 for geographic datasets.

CNN, RNN, or ViT? An Evaluation of Different Deep Learning Architectures for Spatio-Temporal Representation of Sentinel Time Series

Rich information in multitemporal satellite images can facilitate pixel-level land cover classification. However, what is the most suitable deep learning architecture for high-dimension spatio-temporal representation of remote sensing time series remains unclear. In this study, we theoretically analyzed the different mechanisms of the different deep learning structures, including the commonly used convolutional neural network (CNN), the high-dimension CNN [three-dimensional (3-D) CNN], the recurrent neural network, and the newest vision transformer (ViT), with regard to learning and representing the temporal information for spatio-temporal data. The performance of the different models was comprehensively evaluated on large-scale Sentinel-1 and Sentinel-2 time-series images covering the whole of Slovenia. First, the 3-D CNN, long short-term memory (LSTM), and ViT, which all have specific structures that preserve temporal information, can effectively extract the spatio-temporal information, with the 3-D CNN and ViT showing the best performance. Second, the performance of the 2-D CNN, in which the temporal information is collapsed, is lower than that of the 3-D CNN, LSTM, and ViT but outperforms the conventional methods. Thirdly,using both optical and synthetic aperture radar (SAR) images performs almost the same as using only optical images, indicating that the information that can be extracted from optical images is sufficient for land-cover classification. However, when optical images are unavailable, SAR imagescan provide satisfactorily classification results. Finally, the modern deep learning methods can effectively overcome the disadvantages in imaging conditions where parts of an image or images of some periods are missing. The testing data are available at <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><uri>gpcv.whu.edu.cn/data</uri></i> .

Roadway Vulnerability Assessment against Hurricanes Using Satellite Images

Infrastructures such as roadways, power lines, and communications networks play a critical role in our society. However, they are also susceptible to failures, especially those caused by extreme events, quickly affecting large geographical areas. Predicting where and when these failures will occur with high confidence is very difficult because of the stochastic nature of such events. Nevertheless, it is possible to know in advance which areas are more vulnerable and plan accordingly. This paper aims to use remote sensing techniques based on satellite images to detect roadways’ vulnerabilities to hurricanes. Each road is assigned a vulnerability score based on the information retrieved from the satellite image and other geographic information system (GIS) data, which can be used to create a vulnerability map of the area. Both vegetation exposure along the roads and the consequences of the closures are considered in the vulnerability assessment. The study area has been selected in Tallahassee, Florida, U.S., where a high-resolution satellite image was acquired in September 2018, 2 weeks before Hurricane Michael. The findings of this work can help management teams and city responders to identify the most vulnerable regions at risk of disruption and organize resources before the event.

Representing and modeling spatio-temporal uncertainty using belief function theory in flood extent mapping

In our world of rapid change, floods are a growing threat. In this context, flood extent mapping is important for damage and accurate and timely information about flood-affected areas are needed. Due to the suitability of remote sensing in mapping flood inundation, multi-source satellite images have been used in recent years. This paper presents a new approach to representing and modeling uncertainty in spatio-temporal information of satellite images using belief function theory for flooded areas extraction. Interval approaches and the associated arithmetic operations are used for modeling the uncertainty of spatial locations and temporal durations from 2D dimensional model in order to take the direction of flow in the flood. Based on topological spatial relationships, the similarity for the spatio-temporal inundation events is performed and the similar intervals are fused. Detailed examples are given to illustrate each step. A case study of flood risk in Chad Lake from 2003 to 2019 using Sentinel-2, MODIS and Landsat-8 images shows how spatio-temporal uncertainty can be problematic and bias the results of even the simplest flood mapping, and how coupling a Dempster–Shafer approach with topological relationships nevertheless offers a useful solution.

QUantitative and Automatic Atmospheric Correction (QUAAC): Application and Validation.

The difficulty of atmospheric correction based on a radiative transfer model lies in the acquisition of synchronized atmospheric parameters, especially the aerosol optical depth (AOD). At the moment, there is no fully automatic and high-efficiency atmospheric correction method to make full use of the advantages of geostationary meteorological satellites in large-scale and efficient atmospheric monitoring. Therefore, a QUantitative and Automatic Atmospheric Correction (QUAAC) method is proposed which can efficiently correct high-spatial-resolution (HSR) satellite images. QUAAC uses the atmospheric aerosol products of geostationary satellites to match the synchronized AOD according to the temporal and spatial information of HSR satellite images. This method solves the problem that the AOD is difficult to obtain or the accuracy is not high enough to meet the demand of atmospheric correction. By using the obtained atmospheric parameters, atmospheric correction is performed to obtain the surface reflectance (SR). The whole process can achieve fully automatic operation without manual intervention. After QUAAC applied to Gaofen-2 (GF-2) HSR satellite and Himawari-8 (H-8) geostationary satellite, the results show that the effect of QUAAC correction is slightly better than that of the Fast Line-of-sight Atmospheric Analysis of Spectral Hypercubes (FLAASH) correction, and the QUAAC−corrected surface spectral curves have good coherence to that of the synchronously measured by field experiments.

A Deep Unfolded Prior-Aided RPCA Network for Cloud Removal

Clouds, together with their shadows, usually occlude ground-cover features in optical remote sensing images. This hinders the utilization of these images for a range of applications such as earth observation, land-cover classification and urban planning. In this work, we propose a deep unfolded and prior-aided robust principal component analysis (DUPA-RPCA) network for removing clouds and recovering ground-cover information in multi-temporal satellite images. We model these cloud-contaminated images as a sum of low rank and sparse elements and then unfold an iterative RPCA algorithm that has been designed for reweighted <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\ell _{1}$</tex-math></inline-formula> minimization. As a result, the activation function in DUPA-RPCA adapts for every input at each layer of the network. Our experimental results on both Landsat and Sentinel images indicate that our method gives better accuracy and efficiency when compared with existing state of the art methods.

Novel approach of association rule mining for tree canopy assessment

The evolution of technology and availability of voluminous satellite images are bringing a new scenario in satellite image classification where a performance efficient method for predictive analysis of satellite images for land cover classification needs to be devised. As urban areas are growing at faster rate, special attention needs to be given to solve tree canopy assessment problem. Vegetation indices are calculated from spectral information of satellite images. Hundreds of such vegetation indices are available to detect vegetation from a satellite image. The contribution of this paper is designing an improved Apriori algorithm to select optimal number of vegetation indices for tree canopy assessment. In this research, we propose a novel computational approach that allows the improvement of results. It selects optimal combination of vegetation indices and applies principal component analysis on it. It uses a greedy approach based on Apriori algorithm. This study emphasizes on assessment of tree canopy using GPU-enabled environment for performance-efficient assessment. The results achieved, are comparable to state-of-the-art techniques, with an accuracy of 96%. The research has considered 4 years data for Mumbai city of India. This research is useful for Green India Mission of India to assess tree canopy of urban region.

A new fog and low stratus retrieval for tropical South America reveals widespread fog in lowland forests

Fog-driven epiphyte-rich tropical cloud forests were long believed to be restricted to mountainous regions. Recent studies have shown the occurrence of such forest types in the lowlands of French Guiana, where early morning radiation fog was shown to modify the water and energy cycles at the canopy level of the tropical lowland rain forest (Gehrig-Downie et al., 2012; Obregon et al., 2011). Since this newly discovered Tropical Lowland Cloud Forest (TLCF) harbours a unique biodiversity comparable to mountain cloud forests, knowledge of its spatial distribution across tropical lowland forests is of large ecological interest, but so far fully unknown. A prerequisite to detect potential TLCF stands is knowledge on the spatial occurrence of fog/low stratus clouds (FLS) over longer time scales, indicating fog frequency. We address this need on a continental scale by presenting the first spatially-explicit, high-resolution product on nocturnal FLS occurrence over the entire tropical lowland rain forest area of South America. The product is based on a new FLS retrieval scheme that uses brightness temperature (BT) information of night-time satellite images acquired by the Aqua Moderate Imaging Spectroradiometer (Aqua MODIS). Since landforms substantially influence small-scale FLS development, we used a dynamic threshold technique depending on the spatially variable subpixel information on landform suitability for FLS generation. The product was successfully validated against in-situ visibility measurements combined with a plausibility analysis based on the published but scarce fog observations over the Amazon. The independence of the new algorithm from measured total precipitable water (TPW) is an improvement over existing procedures that are limited to the availability of these ancillary data. The FLS frequency map for the entire tropical lowland forest of South America derived from over 52,000 scenes (18 years from 2003 to 2020 inclusive) indicates that FLS is widespread, though spatially differentiated, throughout the tropical lowland forests of South America, providing suitable environmental conditions for the TLCF. This knowledge on the spatial distribution of potential TLCF is of major importance for conservation efforts. The new FLS scheme is applicable to all tropical lowland forests and can therefore support global conservation efforts of the valuable TLCF ecosystems.

Canopy Temperature as a Key Physiological Trait to Improve Yield Prediction under Water Restrictions in Potato

Canopy temperature (CT) as a surrogate of stomatal conductance has been highlighted as an essential physiological indicator for optimizing irrigation timing in potatoes. However, assessing how this trait could help improve yield prediction will help develop future decision support tools. In this study, the incorporation of CT minus air temperature (dT) in a simple ecophysiological model was analyzed in three trials between 2017 and 2018, testing three water treatments under drip (DI) and furrow (FI) irrigations. Water treatments consisted of control (irrigated until field capacity) and two-timing irrigation based on physiological thresholds (CT and stomatal conductance). Two model perspectives were implemented based on soil water balance (P1) and using dT as the penalizing factor (P2), affecting the biomass dynamics and radiation use efficiency parameters. One of the trials was used for model calibration and the other two for validation. Statistical indicators of the model performance determined a better yield prediction at harvest for P2, especially under maximum stress conditions. The P1 and P2 perspectives showed their highest coefficient of determination (R2) and lowest root-mean-squared error (RMSE) under DI and FI, respectively. In the future, the incorporation of CT combining low-cost infrared devices/sensors with spatial crop models, satellite image information, and telemetry technologies, an adequate decision support system could be implemented for water requirement determination and yield prediction in potatoes.

Multi-spectral cloud detection based on a multi-dimensional and multi-grained dense cascade forest

Cloud detection in satellite images is a vital step for cloud/land recognition, cloud/snow discrimination, and cloud shadow removal. Accurate cloud detection plays an important role in land resource management, environmental pollution monitoring, and land target recognition. Deep learning (DL) algorithms have shown great progress in cloud detection. However, as the complexity of the DL-based model increases, cloud detection efficiency decreases. DL-based cloud detection models are unable to successfully balance the performance-efficiency tradeoff. In our study, a multi-dimensional and multi-grained dense cascade forest (MDForest) is proposed for multi-spectral cloud detection. MDForest is a deep forest structure that automatically extracts low-level and high-level features from satellite cloud images end-to-end; a multi-dimensional and multi-grained scanning mechanism is introduced to capture the spectral information of multi-spectral satellite images while enhancing the representation learning ability of cascade forest. The experimental results on the HJ-1A/1B dataset show that MDForest improves the performance of cloud detection and possesses a good inference efficiency compared with DL-based cloud detection methods, which makes the proposed MDForest satisfy the application where good performance and high efficiency are both required.

Integration of Texture and Spectral Response with AI Techniques for Buildings Footprint Identification Using High-Resolution Satellite Images

The rapid growth and advancement in the application of remote sensing (RS) and geographic information systems (GIS) in several application domains helped many researchers to analyze a wide range of information from satellite imagery. Settlement information that includes building foot prints is a very essential parameter for different applications such as urban planning, environmental planning and disaster management. VGG-16 (Visual Geometry Group) conventional neural network (CNN) model is a popular model used to detect and classify the input images. In this study, a new approach was proposed by integration of VGG-16-CNN model with spectral and textural information of satellite images for identification of building foot prints. The model was trained and implemented to identify the building footprints using Worldview-3 high-resolution satellite image over part of Mumbai city of Maharashtra state, India. Classification accuracy in the proposed approach is observed to be nearly 94% as compared to 82% in case of single-shot detector (SSD) algorithm alone. Metric parameters such as F1 score of 0.957, intersection over union (IoU) of 94.86% and total error rate of 8.133% also indicated better performance of the proposed approach. Particularly, the approach is highly beneficial for urban development authorities as they need to monitor the large number of vacant lands spread across urban areas.

ASSESMENT OF WATER RESOURCES USING LANDSAT SATELLITE IMAGERY

Observation of surface water is a very important for studying hydrological processes and last advances in satellite-based optical remote sensors have promoted the field of sensing surface water of a certain area or from a catchment area. This paper has considered analysing data from satellite image and geographical information system (GIS) for a catchment area. Banat hydrographical area is situated in the western part and south-western part of Romania and Barzava River basin drains a relatively small area in Banat Hydrographical Area. GIS techniques were used for extract data NDWI, MNDWI, AWEI indexes from in Landsat-8 satellite in images to evaluate their performances for the extraction of surface water. This paper shows the monitoring of Barzava river basin based on the indices resulting from the processing of satellite images. The objective of this research was to extract the surface water bodies from the hydrographic basin of river Barzava. Then, these data are utilized for purpose finding correlation and regression equations between the reflectance of the satellite image and the water parameters. NDWI being an indicator of plant water stress, it is observed that for 2018 compared to 2019 the basin of the Barzava River was affected by the drought.

Land use and Land Cover Change Analysis, Causes and Effects on Agricultural Productivity, Case of Anlemo District in Hadiya zone, Ethiopia

Change in land use land cover (LULC) is one of the constraints which can influence the development of sustainable agricultural in the study area. The change in LULC can negatively affect the possible use of land and completely lead to soil and vegetation degradation that have an influences on crop productivity. LULC change analysis was conducted in Anlemo District in Hadiya zone, Ethiopia for the period of 2000-2018, using Remote Sensing satellite image and Geographic Information System with field verifications. This was to look at LULC change, its causes and influences on crop productivity in Anlemo District. In this study, LULC maps of 2000, 2010 and 2018, and change maps of 2000-2010 and 2010- 2018 were produced. Results from LULC change analysis revealed an increase in agricultural land from 36.6% in 2000 to 55.764% in 2018. The increase of agricultural land was mainly at the cost of vegetation and grazing land cover change. Vegetation cover decreased from 21.81% in 2000 to 14.601% in 2018. Shrub land area was 38.81% in 2000 that decreased to 19.933% in 2018 and wetland which was 0.817% in 2000 increased to 1.057% in 2018. The study also made known that the main reasons of LULC changes were mainly, expansion of agricultural land and clearing vegetation. Therefore, these urges to initiate mechanized farming system that permits small land holder farmers to obtain more crop products from their small plot of land and to use alternative sources for fuel to reduce the complete reliance of rural community on forest products.

Improved framework of many‐objective evolutionary algorithm to handle cloud detection problem in satellite imagery

Automatic cloud detection algorithm based on supervised learning approach has emerged due to its effectiveness in extracting weather information in satellite images. However, algorithm requires field-expert intervention with huge database of training samples to evaluate its clustering performance. Moreover, lacking in availability of labelled data makes difficult to train the input samples. Therefore, this article puts forward unsupervised many-objective evolutionary clustering technique to discriminate cloudy regions on varying characteristic of underlying surfaces. The study begins with the modification in search capability of θ -NSGA-III optimisation algorithm by incorporating penalised vector angle concept in associate operator. The analysis of proposed approach has been carried out on benchmark many-objective DTLZ test problems, compared against original θ -NSGA-III and NSGA-III algorithms. The proposed modified θ -NSGA-III has been further utilised as clustering technique to solve unsupervised cloud detection problem. Optimal centroid vector for clustering using proposed approach is obtained through modified crossover operator, mutation operator and environmental selection method. Experimental results reveal that proposed approach outperforms comparative many-objective algorithms, MOEA/D and NSGA-III for Landsat 8, MODIS and NOAA satellite images with lower classification average error of 2.44 % in cloud detection for most of the evaluated test cases.

SPATIAL-TEMPORAL DISTRIBUTION OF DROUGHT IN THE WESTERN REGION OF PARAGUAY (2005–2017)

Abstract. The Western Region of Paraguay is naturally dry which makes it vulnerable to almost permanent drought events. Hence, low cost drought monitoring is necessary. Therefore, an index derived from satellite image information was used for this purpose. The Normalized Difference Drought Index (NDDI) was used with the objective of studying the characteristics (temporal and spatial distribution) of drought in the Western Region of Paraguay from years 2005 to 2017 and relate drought point values of NDDI to the land use cover of the study area. Terra satellite’s Moderate Resolution Imaging Spectro-radiometer (MODIS) sensor images were used for the calculation of NDDI. The driest month of all years (July) and the wettest month of all years (December) were averaged to analyze spatial tendencies. The NDDI was contrasted spatially with the Land Use Cover of the Western Region to pinpoint the location of the highest values. The year 2015 had the highest value of 0.9847 in agricultural land use in the Department of Boqueron. The NDDI was a good indicator of drought throughout the region and could be a complement for in-situ measurements.

RETRACTED ARTICLE: Adaptive supervised multi-resolution approach based modeling of performance improvement in satellite image classification

Satellite image classification is a significant piece of utilizations in different fields, for example, horticulture, nature observing, and disaster management. This work is expected to improve the spatial and spectral information of satellite images by using higher request bits of knowledge in a blend with power shade l inundation utilizing Adaptive Supervised Multi-Resolution demonstrating classification approach. The proposed Adaptive Supervised Multi-Resolution based strategy that naturally orders the various regions from spatiotransient remote detecting pictures. Initially, a kernel representation has been planned by the structure of multispectral and temporal remote detecting information. Also, the Adaptive Supervised Multi-Resolution system with tweaked parameters has been proposed for preparing region tests and learning spatiotemporal discriminative representations. The following parameters are used to assess the execution of the proposed Adaptive Supervised Multi-Resolution: sensitivity, specificity, accuracy, and false classification ratio. To reduce the dimensionality of multi-band satellite images, the autonomous part assessment is used, which uses the higher-organize experiences of the data independently. The exhibition of the proposed technique was approved through simulation utilizing the MATLAB programming. Compared with ordinary satellite image classifier with a proposed Adaptive Supervised Multi-Resolution strategy based classifier execution by accomplishing the result of 97.695% in accuracy, 94.815% in sensitivity, and 97.75% in specificity.

Self-Supervised Representation Learning for Remote Sensing Image Change Detection Based on Temporal Prediction

Traditional change detection (CD) methods operate in the simple image domain or hand-crafted features, which has less robustness to the inconsistencies (e.g., brightness and noise distribution, etc.) between bitemporal satellite images. Recently, deep learning techniques have reported compelling performance on robust feature learning. However, generating accurate semantic supervision that reveals real change information in satellite images still remains challenging, especially for manual annotation. To solve this problem, we propose a novel self-supervised representation learning method based on temporal prediction for remote sensing image CD. The main idea of our algorithm is to transform two satellite images into more consistent feature representations through a self-supervised mechanism without semantic supervision and any additional computations. Based on the transformed feature representations, a better difference image (DI) can be obtained, which reduces the propagated error of DI on the final detection result. In the self-supervised mechanism, the network is asked to identify different sample patches between two temporal images, namely, temporal prediction. By designing the network for the temporal prediction task to imitate the discriminator of generative adversarial networks, the distribution-aware feature representations are automatically captured and the result with powerful robustness can be acquired. Experimental results on real remote sensing data sets show the effectiveness and superiority of our method, improving the detection precision up to 0.94–35.49%.

Análise da fragmentação da paisagem na região extremo Oeste Paulista, Brasil

O conceito de estrutura da paisagem se refere ao estudo da paisagem a partir da medição e quantificação de seus aspectos estruturais e espaciais. O objetivo deste artigo é avaliar as métricas da paisagem no extremo Oeste Paulista, Brasil. Para isso, usamos informações de imagens de satélite e um Sistema de Informação Geográfica, a fim de identificar o grau de retalhamento e isolamento da paisagem. Como resultados temos que a área é recoberta por usos antrópicos, que representavam 73,92%, e o restante está distribuído em áreas naturais (área úmida, ilha e fragmentos florestais e corpos d'água). Ao analisarmos o índice de área, referente ao tamanho dos fragmentos, pudemos identificar que os maiores percentuais estão concentrados na pastagem (42,51%) e agricultura (28,58%). Os índices analisados confirmam que a fragmentação da paisagem é expressiva e sofre intensas perturbações.