SPOT Satellite Research Articles

Assessment of Deep Learning Techniques for Land Use Land Cover Classification in Southern New Caledonia

Land use (LU) and land cover (LC) are two complementary pieces of cartographic information used for urban planning and environmental monitoring. In the context of New Caledonia, a biodiversity hotspot, the availability of up-to-date LULC maps is essential to monitor the impact of extreme events such as cyclones and human activities on the environment. With the democratization of satellite data and the development of high-performance deep learning techniques, it is possible to create these data automatically. This work aims at determining the best current deep learning configuration (pixel-wise vs. semantic labelling architectures, data augmentation, image prepossessing, …), to perform LULC mapping in a complex, subtropical environment. For this purpose, a specific data set based on SPOT6 satellite data was created and made available for the scientific community as an LULC benchmark in a tropical, complex environment using five representative areas of New Caledonia labelled by a human operator: four used as training sets, and the fifth as a test set. Several architectures were trained and the resulting classification was compared with a state-of-the-art machine learning technique: XGboost. We also assessed the relevance of popular neo-channels derived from the raw observations in the context of deep learning. The deep learning approach showed comparable results to XGboost for LC detection and over-performed it on the LU detection task (61.45% vs. 51.56% of overall accuracy). Finally, adding LC classification output of the dedicated deep learning architecture to the raw channels input significantly improved the overall accuracy of the deep learning LU classification task (63.61% of overall accuracy). All the data used in this study are available on line for the remote sensing community and for assessing other LULC detection techniques.

Correction of Directional Effects in VEGETATION NDVI Time-Series

Land surface reflectance measurements from the VEGETATION program (SPOT-4, SPOT-5 and PROBA-V satellites) have led to the acquisition of consistent time-series of Normalized Difference Vegetation Index (NDVI) at a global scale. The wide imaging swath (>2000 km) of the family of VEGETATION space-borne sensors ensures a daily coverage of the Earth at the expense of a varying observation and illumination geometries between successive orbit overpasses for a given target located on the ground. Such angular variations infer saw-like patterns on time-series of reflectance and NDVI. The presence of directional effects is not a real issue provided that they can be properly removed, which supposes an appropriate BRDF (Bidirectional Reflectance Distribution Function) sampling as offered by the VEGETATION program. An anisotropy correction supports a better analysis of the temporal shapes and spatial patterns of land surface reflectance values and vegetation indices such as NDVI. Herein we describe a BRDF correction methodology, for the purpose of the Copernicus Global Land Service framework, which includes notably an adaptive data accumulation window and provides uncertainties associated with the NDVI computed with normalized reflectance. Assessing the general performance of the methodology in comparing time-series between normalized and directional NDVI reveals a significant removal of the high-frequency noise due to directional effects. The proposed methodology is computationally efficient to operate at a global scale to deliver a BRDF-corrected NDVI product based on long-term Time-Series of VEGETATION sensor and its follow-on with the Copernicus Sentinel-3 satellite constellation.

A high-contrast search for variability in HR 8799bc with VLT-SPHERE

ABSTRACT The planets HR8799bc display nearly identical colours and spectra as variable young exoplanet analogs such as VHS 1256-1257ABb and PSO J318.5-22, and are likely to be similarly variable. Here we present results from a 5-epoch SPHERE IRDIS broadband-H search for variability in these two planets. HR 8799b aperture photometry and HR 8799bc negative simulated planet photometry share similar trends within uncertainties. Satellite spot lightcurves share the same trends as the planet lightcurves in the August 2018 epochs, but diverge in the October 2017 epochs. We consider Δ(mag)b − Δ(mag)c to trace non-shared variations between the two planets, and rule out non-shared variability in Δ(mag)b − Δ(mag)c to the 10–20 ${{\ \rm per\ cent}}$ level over 4–5 h. To quantify our sensitivity to variability, we simulate variable lightcurves by inserting and retrieving a suite of simulated planets at similar radii from the star as HR 8799bc, but offset in position angle. For HR 8799b, for periods <10 h, we are sensitive to variability with amplitude $\gt 5{{\ \rm per\ cent}}$. For HR 8799c, our sensitivity is limited to variability $\gt 25{{\ \rm per\ cent}}$ for similar periods.

Effect of Erosion and Accretion on Beach Profile in Kuala Terengganu Coastal Areas

Coastal erosion and accretion are long-term process that may cause changes in shoreline and beach profiles. Due to erosion and accretion effects, most of the coastal areas in Malaysia are suffering from destruction of property especially at the coastal areas in east coast of Peninsular Malaysia. This study was conducted to determine the effects of erosion and accretion on beach profiles at four (4) coastal areas in Kuala Terengganu using remote sensing and GPS observation methods. The objectives include to derive the coastal erosion and accretion rate, to measure coastal elevation for beach slope angle calculation, and to determine the relationship between beach slope angle with coastal erosion and accretion. The erosion and accretion rate was derived from SPOT-5 satellite image and unmanned aerial vehicles (UAV). In order to obtain the beach profile, the elevation with 50m offset for every chainage, and 5m offset for each cross-section point were carried out using real-time kinetic (RTK) observation methods. It was found that the highest and lowest erosion and accretion rates were 170.29 m²/year and 57.53 m²/year, respectively. The beach profile became steeper with the beach slope values of 11.004° and 7.652° at high and low erosion areas, respectively. The relationships between beach slope angle and coastal erosion/accretion were found as 0.12 and 0.86 respectively. The findings showed that steeper beach profiles influenced the high rate values of erosion and accretion. For more accurate findings, further studies on the factors affecting the erosion and accretion such as monsoon seasonal changes and morphological impact are necessary to support the reliable decision-making process for sustainable coastal management.

Distribution and area changes of mangroves with remote sensing technology in Perancak Estuary, Jembrana, Bali

Mangrove vegetation has an important role for its functions and benefits to the coastal area and the society which lives in it. The purpose of this research is to know the distribution and the changes of mangrove vegetation area located in Perancak Estuary area within 10 years, and to know the green-ness level and the environmental parameters of mangrove vegetation; salinity and substrate condition. To observe the distribution and the changes of mangrove vegetation area, supervised classification is used on SPOT satellite imagery in 2007 and 2017. The result shows that the expansion of mangrove vegetation area in Perancak Estuary is 18.7 hectares along with the increasing of the green-ness level. In addition, the environmental parameters which are the salinity and the substrate conditions are also observed, where the average salinity value is 31.54 ppt and the substrate conditions are good (not much waste).

Unsupervised Change Detection Using Spectrum-Trend and Shape Similarity Measure

The emergence of very high resolution (VHR) images contributes to big challenges in change detection. It is hard for traditional pixel-level approaches to achieve satisfying performance due to radiometric difference. This work proposes a novel feature descriptor that is based on spectrum-trend and shape context for VHR remote sensing images. The proposed method is mainly composed of two aspects. The spectrum-trend graph is generated first, and then the shape context is applied in order to describe the shape of spectrum-trend. By constructing spectrum-trend graph, spatial and spectral information is integrated effectively. The approach is performed and assessed by QuickBird and SPOT-5 satellite images. The quantitative analysis of comparative experiments proves the effectiveness of the proposed technique in dealing with the radiometric difference and improving the accuracy of change detection. The results indicate that the overall accuracy and robustness are both boosted. Moreover, this work provides a novel viewpoint for discriminating changed and unchanged pixels by comparing the shape similarity of local spectrum-trend.

Investigation of the State Vectors and Prediction of the Orbital Elements for Spot-6 Satellite during 1300 periods with Perturbations

Computer simulations were carried out to investigate the dependence of the main perturbation parameters (Sun and Moon attractions, solar radiation pressure, atmosphere drag, and geopotential of Earth) on the orbital behavior of satellite. In this simulation, the Cowell method for accelerations technique was adopted, the equation of motion with perturbation was solved by 4th order Runge-Kutta method with step (1/50000) of period to obtain the state vectors for position and velocity. The results of this simulation have been compared with data that available on TLEs (NORD data in two line elements). The results of state vectors for satellites (Cartosat-2B, Gsat-14 and Spot-6) shows excellent correlation and this is leading us to extend our study for (spot-6) satellite to include the orbital behavior during 13000 periods under the effect of one type of perturbation or all types. The results indicate that all perturbation have clear effect on spot-6 orbit, reduced the perigee and apogee about 3 Km. during 89 days, also the time of period reduced 4.7 sec. Other conclusions present that the perigee angle increases 28.01 degree with any perturbation accept SRP. Furthermore, the geopotential have a big periodic effect but the atmospheric drag have accumulated effect on most orbital elements.

TOWARDS AN AUTOMATED FLOOD AREA EXTRACTION FROM HIGH RESOLUTION SATELLITE IMAGES

Abstract. Flooding is considered as one of the most devastated natural disasters due to its adverse effect on human lives as well as economy. Since more population concentrate towards flood prone areas and frequent occurrence of flood events due to global climate change, there is an urgent need in remote sensing community for faster and reliable inundation mapping technologies to increase the preparedness of population and reduce the catastrophic impact. With the recent advancement in remote sensing technologies and integration capability of deep learning algorithms with remote sensing data makes faster mapping of large area is feasible. Therefore, this study attempted to explore a faster and low cost solution for flood area extraction by integrating convolution neural networks (CNNs) with high resolution (1.5 m) SPOT satellite images. By consider the system requirement as a measure of cost, capabilities (speed and accuracy) of a deeper (ResNet101) and a shallower (MobileNetV2) CNNs on flood mapping were examined and compared. The models were trained and tested with satellite images captured during several flood events occurred in Japan. It is observed from the results that ResNet101 obtained better flood area mapping accuracy than MobileNetV2. Whereas, MobileNetV2 is having much higher capabilities in faster mapping in 0.3 s/km2 with a competitive accuracy and minimal system requirements than ResNet101.

Modelling land use/land cover changes prediction using multi-layer perceptron neural network (MLPNN): a case study in Makassar City, Indonesia

ABSTRACT This study used Remote Sensing and Geographic Information System (GIS) tools to produce a predictive model of land use/land cover changes in Makassar City by 2031. The model was based on the classification of SPOT satellite images from 2006, 2011, and 2016 with an assessment of 10 driving factors. The Multi-Layer Perceptron Neural Network (MLPNN) method was run with the Markov Chain model using the Business as Usual scenario. The results predict that built-up areas will cover about 80% of the total area of Makassar City by 2031. This research can inform stakeholder decision-making in Makassar City.

Combining and Comparing an Unmanned Aerial Vehicle and Multiple Remote Sensing Satellites to Calculate Long-Term River Discharge in an Ungauged Water Source Region on the Tibetan Plateau

Research into global water resources is challenged by the lack of ground-based hydrometric stations and limited data sharing. It is difficult to collect good quality, long-term information about river discharges in ungauged regions. Herein, an approach was developed to determine the river discharges of 24 rivers in ungauged regions on the Tibetan Plateau on a long-term scale. This method involved coupling the Manning–Strickler formula, and data from an unmanned aerial vehicle (UAV) and the Gaofen-2, SPOT-5, and Sentinel-2 satellites. We also compared the discharges calculated by using the three satellites’ data. Fundamental information about the rivers was extracted from the UAV data. Comparison of the discharges calculated from the in-situ measurements and the UAV data gave an R2 value of 0.84, an average NSE of 0.79, and an RMSE of 0.11 m3/s. The river discharges calculated with the GF-2 remote sensing data and the in-situ experiments for the same months were compared and the R2, RMSE, and the NSE were 0.80, 1.8 m3/s, and 0.78, respectively. Comparing the discharges calculated over the long term from the measured in-situ data and the SPOT-5 and Sentinel-2 data gave R2 values of 0.93 and 0.92, and RMSE values of 2.56 m3/s and 3.16 m3/s, respectively. The results showed that the GF-2 and UAV were useful for calculating the discharges for low-flow rivers, while the SPOT-5 or the Sentinel-2 satellite gave good results for high-flow river discharges in the long-term. Our results demonstrate that the discharges in ungauged tributaries can be reliably estimated in the long-term with this method. This method extended the previous research, which described river discharge only in one period and provided more support to the monitoring and management of the tributaries in ungauged regions.

Fragmentation dynamics in an Abies religiosa forest of central Mexico

Fir (Abies Mill.) forests of Mexico are relicts of the boreal forests that advanced southwards during glaciation periods. Mexico is a center of diversification of the Abies genus, as there are eight species in its territory, six of which are endemic. The forests of Abies religiosa (Kunth) Schltdl. & Cham. near Mexico City are subject to a process of deterioration. We analyzed the fragmentation dynamics of the A. religiosa forest in the northern region of the Sierra Nevada, Mexico. Land cover change detection was done by means of high-resolution images acquired by the SPOT satellite in 2005, 2010, 2015, and 2018. Habitat fragmentation was observed, with a decrease in the size of dense Abies masses. The area covered by Abies decreased by 22.9%. The area occupied by forest openings increased 3% from 2005 to 2010 and then decreased by 1.8% and 1.6% in the following periods. The land patch type Other Forest Cover increased in both frequency and size, with the area increasing by 23.3%, which warns of a change process towards this patch type. The formation of increasingly smaller and isolated remnants of A. religiosa forest in the Sierra Nevada can lead to the loss of this vegetation relict and its replacement by other types of cover in the short term.

Estimation of daily CO2 fluxes and of the components of the carbon budget for winter wheat by the assimilation of Sentinel 2-like remote sensing data into a crop model

Croplands contribute to greenhouse gas emissions but also have the potential to mitigate climate change through soil carbon storage. However, there is a lack of tools based on objective observations for assessing cropland C budgets at the plot scale over large areas. Such tools would allow us to more precisely establish the contribution of an agricultural plot to net CO2 emissions according to the plot management and identify levers for improving the C budget. In this study, we present a diagnostic regional modelling approach, called SAFY-CO2, that assimilates high spatial and temporal resolution (HSTR) optical remote sensing data in a simple crop model and evaluate the performance of this approach in quantifying crop production and the main components of the annual carbon budget for winter wheat.The SAFY-CO2 model simulates daily crop development (biomass, partition to leaves, etc.), the components of net ecosystem CO2 fluxes, and the annual yield and net ecosystem carbon budget (NECB).Multi-temporal green area index (GAI) maps derived from HSTR data from the Formosat-2 and SPOT satellites were used to calibrate the light-use efficiency and phenological parameters of the model. Data from the literature were used to set a priori values for a set of model parameters, and a large dataset of in situ data was used for model validation. This dataset includes 8 years of eddy-covariance net CO2 flux measurements and GAI, biomass and yield data acquired at 2 instrumented sites in southwest France. Biomass and yield data from 16 fields in the study area between 2005 and 2014 were also used for validation.The SAFY-CO2 model is able to reproduce both GAI dynamics (RRMSE = 14%, R2 = 0.97) and biomass production and yield (RRMSE of 27% and 21%, respectively) with high precisions under contrasting climatic, environmental and management conditions. Additionally, the net CO2 flux components estimated by the model generally agreed well with in situ data and presented very good and significant correlations (RMSE of 1.74, 1.13 and 1.29 gC.m−2.d-1 for GPP, Reco and NEE, respectively; R2 of 0.90, 0.75 and 0.85 for GPP, Reco and NEE, respectively) over the 8 studied years. This study also highlights the importance of accounting for post-harvest vegetative events (spontaneous re-growth, weed development and cover crops) for an accurate calculation of the annual net CO2 flux. This approach requires a limited number of input parameters for estimating yield and net CO2 flux components, which is promising for regional/global-scale applications based on Sentinel 2-like data; however, the approach requires plot-scale data concerning organic amendments and straw management (exportation) in animal farming systems to calculate field C budgets.

Dual-Band Reflectarray to Generate Two Spaced Beams in Orthogonal Circular Polarization by Variable Rotation Technique

This contribution describes the design of a reflectarray to produce two adjacent beams per feed in orthogonal circular polarization (CP) simultaneously at transmit (Tx) and receive (Rx) frequencies in the Ka band. The discrimination in CP is achieved by the variable rotation technique (VRT), whose dual band operation is validated for the first time by the manufacture and measurement of a 25 cm × 25 cm reflectarray demonstrator. Moreover, an optimization procedure has been implemented to improve the narrowband performance of the VRT (which leads to high cross-polar levels when the VRT is applied). The reflectarray prototype has been designed to deviate by ±5° the beams radiated by a dual-CP feed-horn at Tx and Rx frequencies in the Ka band (20/30 GHz). The results are satisfactory and validate the concept of generating two spaced beams in orthogonal CP and dual band by the VRT with a single feed. This concept can be suitable for multiple spot satellites in the Ka band, enabling a reduction in the number of antennas and feeds needed to provide the multispot coverage.

Spatial and temporal evolution of co-seismic landslides after the 2005 Kashmir earthquake

Large earthquakes in mountainous regions often trigger widespread landslides, some of which remain active for years to decades. This study has evaluated the spatial and temporal evolution of the 2005 Kashmir earthquake-induced landslides using pre-earthquake and post-earthquake landslide inventories. Temporal landslide inventories were developed from the SPOT satellite images, using the visual image interpretation supported with field validation. It is observed that after the earthquake, the landslide area and number showed a surge, however, decays in the subsequent years. The rate of decline in the landslide area was relatively slower until 2010, however, it dropped considerably during 2010–2018. The spatial variation of the landslides was traced and analyzed to classify them based on the activity. The temporal evolution of the landslides was compared with the geology, earthquake triggering fault and topographic attributes of slope, aspect, and elevation, to evaluate their influence on the distribution of landslides. The landslide inventories were related to the major triggering factors including rainfall and earthquakes, to determine their impact on the temporal dynamics of the landslides. Regeneration of vegetation is analyzed using the temporal NDVI images as one of the drivers for the decline in the landslide area. The knowledge of temporal dynamics of landslides and their major contributing factors shall assist to understand their triggering mechanism and effective mitigation measures.

Spatial Analysis of Sustainable Land Use Development Coastal Areas in Batang Regency

Batang Regency as the northern coastal area of Java islandexperiences a problem of land use management that has not been optimal so that there is a need for analysis of land use development in the coastal area of Batang Regency. The purpose of this study is to examine the direction of land use development in coastal areas of Batang Regency with spatial analysis. This research method implementsquantitative approach while the analytical techniqueis spatial analysis. Spatial analysis is performed to identify land use of thecoastal area in Batang Regency. The study scope of coastal areasis based on administrative subdistricts located along the coast of Batang Regency. This research utilizes Spot satellite image year 2015 with the map of regionalspatial pattern plan of Batang Regency in 20112031. The spatial analysis result of land use developmentinBatang Regency’s coastal area includes the suitable category of 18.130,65 (56,32%) while the unsuitable category is 14.059,44 (43,68%). Batang Sub-district’s suitable category is 1.214,03 ha, while the unsuitable category is 2.220,51 ha. Kandeman Sub-district’s suitable category is 1.565.02 ha, while the unsuitable category is 2.610,65 ha. TulisSub-district’ssuitablecategory is 2.452,98 ha, while the unsuitable category is 2.055,80 ha. Subah Sub-district’ssuitablecategory is 5.381,89 ha, while the unsuitable category is 2.970,28 ha. Banyuputih Sub-district’s suitable category is 2.314,61 ha, while the unsuitable category is 2.127,89 ha. Gringsing Subdistrict’s suitablecategory is 5.202,12 ha, while the unsuitable categoryis 2.074,31 ha.The conclusion of land use planning development for Batang and Kandeman Sub-districts is that there are still a lot of lands designated for residential development. Subah Sub-district is for fisheries development while Tulis, Banyuputih, and Gringsing Sub-districts are for industrial development.

Glacial geomorphology of the Strait of Magellan ice lobe, southernmost Patagonia, South America

ABSTRACT We present a geomorphic map of the landforms created by the Patagonian Ice Sheet during the local Last Glacial Maximum and perhaps prior glaciations in southernmost Patagonia. Building on prior work, the new mapping focuses in unprecedented detail on the right lateral and frontal landforms formed by the Strait of Magellan ice lobe. We produced the map using aerial orthophotography, Sentinel-2 and SPOT satellite imagery, ALOS PALSAR digital elevation model and fieldwork to ground-truth preliminary interpretations. We delineate at least five glacial events defined by a sequence of moraine drifts and associated glaciofluvial features. In contrast to previous studies, we propose the Magellan ice lobe extended ∼65 km farther, to the Primera Angostura peninsula, during the local Last Glacial Maximum. Our study provides a new context to establish a precise glacial chronology of the Magellan ice lobe during the last glacial cycle in the middle-high latitudes of the Southern Hemisphere.

Analysing suitability of arable land within a convertible production forest (HPK)

Potential land resources become one basic information needed in planning agroecological system development. The purpose of this study was to analyse the suitability of arable land in a convertible production forest area (HPK) in Konawe Kepulauan Regency, Southeast Sulawesi Province. Identification of arable land types and determination of agroecological zones was using spatial analysis through SPOT satellite imagery processing and slope map overlays and soil types to determine arable land units. Existing plant species were recorded and adjusted to the direction of land use from the Agricultural Research and Development Agency. From this study, 30 units of fertile land were obtained covering an area of 12,873.69 ha and spread across six zones, namely zone I covering 182.95 ha (1.42%), II covering 3,102.10 ha (24.10%), III covering 3,125.83 ha (24.28%), IV covering 3,594.36 ha (27.92%), V covering 631.85 ha (4.91%) and VI covering 2,236.61 ha (17.37%) with the distribution of plant species in the form of natural forests, annual crop agriculture, seasonal crops, agroforestry crops and food crops. Plants in zones I, II and III in accordance with the direction of land use but Zones IV, V and VI still not suitable.

Efficient Hybrid Method for Pan-Sharpening Enhancement of Multiband Satellite Images

Pan-sharpening considers one of the most important applications for satellite images as it enhances spectral and spatial information for the images. Empirical Mode Decomposition (EMD) is one of the most powerful techniques for pan-sharpening. It first decomposes the image into a set of Intrinsic Mode Functions (IMFs) and a residual component. These panchromatic and multispectral components are then fused to create an enhanced pan-sharpened image. This paper presents an efficient hybrid method for enhancing pan-sharpening of multiband images transmitted from satellite to ground stations. The proposed approach combines this EMD technique with the most powerful conventional method; Discrete Wavelet Transform (DWT), to maximize the pan-sharpening gain. The proposed hybrid method is validated using satellite images of Nile Valley and Suez Canal region, Egypt, captured by Spot-4 and Landsat-8 satellites. The results imply that the proposed hybrid method provides better qualitative and quantitative quality comparing with the individual and the most common pan-sharpening methods.

Could land cover influence Aedes aegypti mosquito populations?

Aedes aegypti is mainly found in urban environments, where human activity guarantees the permanent availability of potential larval habitats. The present study aimed to test for a possible association between Ae. aegypti infestation and land cover classes. From 23 February to 22 May 2015, immature mosquitoes (except eggs) were sampled in artificial containers identified in dwellings in Córdoba city, Argentina. The proportion of each land cover class was determined by SPOT-5 (Satellites Pour 1'Observation de la Terre or Earth-observing Satellites) image classification. Generalized linear models were developed to assess a suite of predetermined hypotheses and identified cover class variables associated with Ae. aegypti infestation. Arboreal vegetation was identified as the land cover with the greatest relative importance, negatively associated with Ae. aegypti mosquitoes. More infestation was find in areas with less arboreal vegetation, which corresponds to more urbanized areas.

Where are the avalanches? Rapid SPOT6 satellite data acquisition to map an extreme avalanche period over the Swiss Alps

Abstract. Accurate and timely information on avalanche occurrence is key for avalanche warning, crisis management and avalanche documentation. Today such information is mainly available at isolated locations provided by observers in the field. The achieved reliability, considering accuracy, completeness and reliability of the reported avalanche events, is limited. In this study we present the spatially continuous mapping of a large avalanche period in January 2018 covering the majority of the Swiss Alps (12 500 km2). We tested different satellite sensors available for rapid mapping during the first avalanche period. Based on these experiences, we tasked SPOT6 and SPOT7 for data acquisition to cover the second, much larger avalanche period. We manually mapped the outlines of 18 737 individual avalanche events, applying image enhancement techniques to analyze regions in the shade as well as in brightly illuminated ones. The resulting dataset of mapped avalanche outlines, having unique completeness and reliability, is evaluated to produce maps of avalanche occurrence and avalanche size. We validated the mapping of the avalanche outlines using photographs acquired from helicopters just after the avalanche period. This study demonstrates the applicability of optical, very high spatial resolution satellite data to map an exceptional avalanche period with very high completeness, accuracy and reliability over a large region. The generated avalanche data are of great value in validating avalanche bulletins, in completing existing avalanche databases and for research applications by enabling meaningful statistics on important avalanche parameters.