Open-source Satellite Imagery Research Articles

Dual level attention based lightweight vision transformer for streambed land use change classification using remote sensing

Due to rapid urbanization, rising food demand, and changed precipitation patterns, the waterbodies are contracting their former beds. The continuous shrinking of waterbodies is deteriorating the vital cultural, supporting, provisioning, and regulating services. Thus, understanding and mitigating the impacts of streambed land cover change is crucial for maintaining healthy aquatic ecosystems and improving flood resilience of surrounding population. The existing works use high-resolution aerial imagery focusing on large waterbodies, while ignoring the most vulnerable floodplains of innumerous small water bodies due to high inter-class similarity. This limits the ability to perform a temporal analysis of land cover change along small water bodies. The present work aims to resolve this issue using open-source satellite imagery and taking patched samples along the boundary of small water bodies to identify long-term changes in land cover patterns. Sentinel-2 and Landsat 50 acquired satellite images were used to identify the land cover of this colonized stream bed. The data of Landsat 50 served as historical reference for identifying the changed land use. To capture spatial hierarchies in satellite images effectively, in this paper, a novel dual attention-based vision transformer has been developed for land-cover classification in four categories namely, water, built-up, siltation, and vegetation. The developed model is trained on the data collected from three potential sites in India. The experimental results are validated against seven state-of-the-art deep learning models. The results reveal that the proposed method outperformed all the considered methods by achieving accuracy and precision of 88.4% and 88.9%, respectively, while consuming the least number of parameters. The results reaffirm the concretization and erosion of nature’s flood buffers for economic advancement.

Cliff dynamics in western Crimea

The paper considers the long-term dynamics of clayey cliffs of the Western Crimea in view of the problem of further recreational development of the Crimean Peninsula. Open-source satellite imagery and long-term (over 40 years) cliff section measurements were analyzed. The paper provides data on the geomorphology of individual coastal sites. It is shown that landslides are the main mechanism causing the cliff retreat on the larger part of the coastline. Slumps are typical for the southern part of the region and their movement episodes are relatively rare. Landslides become more active during the winter — spring period, when the moisture content of clayey cliff rocks increases substantially and the abrasion intensifies. Landslides can also be triggered by short-term heavy precipitation, which is usually observed in the summer. There is no definite relationship between the amount of annual precipitation, storm activity and landslide activity, either synchronous or with a time lag. No regularity in landslide dynamics was identified. It was found that the average long-term rates of cliff edge retreat are 0.1–1.2 m/year. In the northern part of the coast, the rates are the highest, decreasing towards the southern part of the region. The obtained average annual rates of cliff edge retreat are significantly less than those previously reported in the literature. There is a decrease in the beach-forming sediment supply due to a reduction in the stretch of the cliffs. Out of 50 km coastline with cliffs, only 39 km are left in the study area due to being covered by various structures and terracing. The paper also discusses anthropogenic activity, which leads to the formation and movement of man-made landslides and an increase in natural landslide activity. It is shown that the construction of transverse beach-retaining structures leads to blocking of littoral sediment transport, and to increase in the cliff retreat rate outside the protected reach.

MINDORO OIL SPILL PROGRESS MONITORING USING SATELLITE IMAGERY

Abstract. Oil spills in the sea pose a significant threat to marine and coastal ecosystems including corals, sea grasses, and mangroves. Responding to oil spills in the high seas can be challenging and costly, as it is difficult to survey the area using ships or planes. Timely monitoring of extents and direction of drift is costly especially when large area is involved. This study used satellite imagery to monitor the progression of an oil spill in a timely manner. Images captured by various types of sensors was used to provide an almost daily oil extent map. Combining opensource satellite imagery, contributed imagery by commercial providers, and imagery from various national space agencies enabled situational awareness during the oil spill incident. Optical and SAR images were processed by using unsupervised K-means classification, for images that undergone image segmentation, and supervised support vector machine classification. The process was used to produce 46 oil extent maps from 196 satellite images from various sources through Sentinel Asia and The International Charter Space and Major Disasters.

Artificial intelligence convolutional neural networks map giant kelp forests from satellite imagery

Climate change is producing shifts in the distribution and abundance of marine species. Such is the case of kelp forests, important marine ecosystem-structuring species whose distributional range limits have been shifting worldwide. Synthesizing long-term time series of kelp forest observations is therefore vital for understanding the drivers shaping ecosystem dynamics and for predicting responses to ongoing and future climate changes. Traditional methods of mapping kelp from satellite imagery are time-consuming and expensive, as they require high amount of human effort for image processing and algorithm optimization. Here we propose the use of mask region-based convolutional neural networks (Mask R-CNN) to automatically assimilate data from open-source satellite imagery (Landsat Thematic Mapper) and detect kelp forest canopy cover. The analyses focused on the giant kelp Macrocystis pyrifera along the shorelines of southern California and Baja California in the northeastern Pacific. Model hyper-parameterization was tuned through cross-validation procedures testing the effect of data augmentation, and different learning rates and anchor sizes. The optimal model detected kelp forests with high performance and low levels of overprediction (Jaccard’s index: 0.87 ± 0.07; Dice index: 0.93 ± 0.04; over prediction: 0.06) and allowed reconstructing a time series of 32 years in Baja California (Mexico), a region known for its high variability in kelp owing to El Niño events. The proposed framework based on Mask R-CNN now joins the list of cost-efficient tools for long-term marine ecological monitoring, facilitating well-informed biodiversity conservation, management and decision making.

Large-scale individual building extraction from open-source satellite imagery via super-resolution-based instance segmentation approach

Building footprint is a primary dataset of an urban geographic information system (GIS) database. Therefore, it is essential to establish a robust and automated framework for large-scale building extraction. However, the characteristic of remote sensing images complicates the application of the instance segmentation method based on the Mask R-CNN model, which ought to be improved toward extracting and fusing multi-scale features. Moreover, open-source satellite image datasets with wider spatial coverage and temporal resolution than high-resolution images may exhibit different coloration and resolution. This study proposes a large-scale building extraction framework based on super-resolution (SR) and instance segmentation using a relatively lower-resolution (>0.6 m) open-sourced dataset. The framework comprises four steps: color normalization and image super-resolution, scene classification, building extraction, and scene mosaicking. We took Hyogo Prefecture, Japan (19,187 km2) as a test area and extracted 1,726,006 (29.12 km2) of the 3,301,488 buildings (32.46 km2), where the number of buildings and footprint area increased by 3.0 % and 5.0 % respectively. The result indicated that the color normalization and image super-resolution could improve the visual quality of open-source satellite images and contribute to building extraction accuracy. Moreover, the improved Mask R-CNN based on Multi-Path Vision Transformer (MPViT) backbone achieved F1 scores of 0.71, 0.70, 0.81, and 0.67 for non-built-up, rural, suburban, and urban areas, respectively, which is better than those of the baseline model and other mainstream instance segmentation approaches. This study demonstrates the potential of acquiring acceptable building footprint maps from open-source satellite images, which has significant practical implications.

Following the herds? A new distribution of hunting kites in Southwest Asia

Remote-sensing analysis of open-source satellite imagery has identified a major, new distribution of undocumented hunting kite structures in northern Arabia. This new data has important implications on the environmental viability of hunting and on possible settlement patterns during the early and middle Holocene. Running across the eastern side of the Nafud Desert in Saudi Arabia, this research has identified star-shaped kites in a distribution that continues on to southern Iraq. From a broader perspective, this new distribution appears to represent a continuation of the well-known arc of kites recorded running principally through southern Syria and eastern Jordan. As well as representing an important archaeological identification in its own right, this new distribution also has important implications in terms of the paleoenvironment of the region, faunal dispersals and human cultural connections.

Identification of Embodied Environmental Attributes of Construction in Metropolitan and Growth Region of Melbourne, Australia to Support Urban Planning

As growth regions evolve to accommodate the increasing population, they need to develop a wider variety of residential properties to accommodate the varying needs of the residents. As a result, the new accommodation is denser which involves higher embodied water carbon and energy. This research compares the construction differences in metropolitan and growth regions of Melbourne to identify embodied carbon, water, and energy. Representative areas of 25 km2 are selected from both regions. The growth region has 80% of the built area comprised of 2nd generation low-rise residential buildings whereas the prolific construction type in the Metropolitan region is mixed purpose industrial with 30% of the built area comprising of this type. The methodology implies open-source satellite imagery to build a spatial dataset in QGIS. The visual identification of the constructions in the study areas enables to identity the materials used in their construction. The total embodied carbon, water, and energy for the Metropolitan region are 32,895 tonnes, 4192 mL, and 3,694,412 GJ, respectively, whereas in the growth region, the totals are 179,376 tonnes carbon, 2533 mL water, and 2,243,571 GJ. Whilst Metropolitan has a significantly higher overall footprint when this is compared to the population of each region, it is shown that the growth region with its current construction type has a higher embodied carbon, water, and energy per head. The total per head for Metropolitan is 226.7 GJ energy, 257 kL water, and 20 tonnes carbon, whereas in the growth region, the embodied energy, water, and carbon, respectively, per head is 287.4 GJ, 324.6 kL, and 22 tonnes. The current performance per head of the growth region is considerably lower than that of Metropolitan. Using diverse residential construction types and efficient materials can serve the demanding needs of denser populated areas.

Fusing Earth observation and socioeconomic data to increase the transferability of large-scale urban land use classification

Monitoring and understanding urban development requires up-to-date information on multiple urban land-use classes. Manual classification and deep learning approaches based on very-high resolution imagery have been applied successfully, but the required resources limits their capacity to map urban land use at larger scales. Here, we use a combination of open-source satellite imagery, constituting of data from Sentinel-1 and Sentinel-2, and socioeconomic data, constituting of points-of-interest and spatial metrics from road networks to classify urban land-use at a national scale, using a deep learning approach. A related challenge for large-scale mapping is the availability of ground truth data. Therefore, we focus our analysis on the transferability of our classification approach, using ground truth labels from a nationwide land-use dataset for the Netherlands. By dividing the country into four regions, we tested whether a combination of satellite data and socioeconomic data increases the transferability of the classification approach, compared to using satellite data only. The results indicate that socioeconomic data increases the overall accuracy of the classification for the Netherlands by 3 percentage points. In a transfer learning approach we find that adding socioeconomic data increases the accuracy between 3 and 5 percentage points when trained on three regions and tested on the independent fourth one. In the case of training and testing on one region and testing on another, the increase in overall accuracy increased up to 9 percentage points. In addition, we find that our deep learning approach consistently outperforms a random forest model, used here as benchmark, in all of the abovementioned experiments. Overall, we find that socioeconomic data increases the accuracy of urban land use classification, but variations between experiments are large.

MDIR Monthly Ignition Risk Maps, an Integrated Open-Source Strategy for Wildfire Prevention

Countries unaccustomed to wildfires are currently experiencing wildfire as a new climate-change reality. Understanding how fire ignition and propagation are correlated with temperature, orography, humidity, wind, and the mixture and age of individual plants must be considered when designing prevention strategies. While wildfire prevention focuses on fire ignition avoidance, firefighting success depends on early ignition detection, meaning that, in either case, ignition plays a major role. The current case study considered three Portuguese municipalities that annually observe frequent fire ignitions (Tomar, Ourém, and Ferreira do Zêzere) as the testing ground for the Modernized Dynamic Ignition Risk (MDIR) strategy, thus evaluating the efficiency of MDIR and the efficacy of the variables used. This methodology uses geographic information systems technology sustained by open-source satellite imagery, along with the Habitat Risk Assessment model from the InVEST software package, as drivers for the MDIR application. The MDIR approach grants frequent update capabilities and fully open-sourced high ignition risk area identification, producing monthly ignition risk maps. The advantage of using this method is the ease of adaptation to any current monitoring strategy, awarding further efficiency and efficacy in reducing ignitions. The approach delivered adequate results in estimating ignitions for the three Portuguese municipalities, achieving, for several months, prediction accuracy percentages of over 70%. For the studied area, MDIR clearly identifies areas of high ignition risk and delivers an average of 62% success in predicting ignitions, thus showing potential for analyzing the impact of policy implementation and monitoring through the strategy design.

Detecting Change at Archaeological Sites in North Africa Using Open-Source Satellite Imagery

Our paper presents a remote sensing workflow for identifying modern activities that threaten archaeological sites, developed as part of the work of the Endangered Archaeology of the Middle East and North Africa (EAMENA) project. We use open-source Sentinel-2 satellite imagery and the free tool Google Earth Engine to run a per-pixel change detection to make the methods and data as accessible as possible for heritage professionals. We apply this and perform validation at two case studies, the Aswan and Kom-Ombo area in Egypt, and the Jufra oases in Libya, with an overall accuracy of the results ranging from 85–91%. Human activities, such as construction, agriculture, rubbish dumping and natural processes were successfully detected at archaeological sites by the algorithm, allowing these sites to be prioritised for recording. A few instances of change too small to be detected by Sentinel-2 were missed, and false positives were caused by registration errors, shadow and movements of sand. This paper shows that the expansion of agricultural and urban areas particularly threatens the survival of archaeological sites, but our extensive online database of archaeological sites and programme of training courses places us in a unique position to make our methods widely available.

Green Apartheid: Urban green infrastructure remains unequally distributed across income and race geographies in South Africa

Urban green infrastructure provides ecosystem services that are essential to human wellbeing. A dearth of national-scale assessments in the Global South has precluded the ability to explore how political regimes, such as the forced racial segregation in South Africa during and after Apartheid, have influenced the extent of and access to green infrastructure over time. We investigate whether there are disparities in green infrastructure distributions across race and income geographies in urban South Africa. Using open-source satellite imagery and geographic information, along with national census statistics, we find that public and private green infrastructure is more abundant, accessible, greener and more treed in high-income relative to low-income areas, and in areas where previously advantaged racial groups (i.e. White citizens) reside. Areas with White residents report 6-fold higher income, have 11.7% greater tree cover, 8.9% higher vegetation greenness and live 700 m closer to a public park than areas with predominantly Black African, Indian, and Coloured residents. The inequity in neighborhood greenness levels has been maintained (for Indian and Coloured areas) and further entrenched (for Black African areas) since the end of Apartheid in 1994 across the country. We also find that these spatial inequities are mirrored in both private (gardens) and public (street verges, parks, green belts) spaces, hinting at the failure of governance structures to plan for and implement urban greening initiatives. By leveraging open-access satellite data and methods presented here, there is scope for civil society to monitor urban green infrastructure over time and thereby hold governments accountable to addressing environmental justice imperatives in the future. Interact with the data here: green-apartheid.zsv.co.za.

Assessing the effect of fire severity on sediment connectivity in central Chile

Chilean territory is recurrently affected by severe wildfires, which drastically reduce the forest cover and promote runoff, soil erosion and slope instabilities. To understand how the geomorphic system responds to wildfires in terms of sediment dynamics, the assessment of sediment connectivity, i.e. the property describing the relationships between compartments of a geomorphic system, is crucial. This study aims to quantify the spatial linkages between fire severity and sediment connectivity to identify common patterns and driving factors. The compound use of field data and open-source satellite imagery helped to apply the Relative differenced Normalized Burn Ratio (RdNBR) and the Index of Connectivity (IC) in the context of two consecutive wildfires, occurred in 2002 and 2015, in the Rio Toro catchment (Chile). The fire severity assessment showed that the 2002 event affected 90% of the catchment, with high severity areas representing around 70%. The 2015 wildfire instead, affected 76% of the catchment with moderate severity around 42%. Accordingly, as result of the sudden reduction in forest cover in severely affected areas, the IC changed after both wildfires with an overall increase of 1.07 and 0.54, respectively. However, only for the second disturbance, it was possible to observe a clear relationship between the RdNBR and the IC variations. The different degree of vegetation cover heterogeneity between the two pre-wildfire scenarios contributed to different fire severity and IC variability between the two disturbances. The use of open-source satellite data and the development of a weighting factor (W), to be used in IC and able to capture the land cover change driven by the wildfires, could make the application of this approach straightforward, promoting its reproducibility in other catchments for land management and risk mitigation purposes.

Using open-source data to explore distribution of built environment characteristics across Kerala, India.

Built environment characteristics in the neighborhood are of utmost priority for a healthy lifestyle in the fast-urbanizing countries. These characteristics are closely linked to the disease burden and challenges in low- and middle-income countries (LMICs), which have been unexplored using open-source data. The present technology offers online resources and open source software that enable researchers to explore built environment characteristics with health and allied phenomena. This article intends to delineate methods to capture available and accessible objective built environment variables for a state in India and determine their distribution across the state. Built environment variables such as population density and residential density were collated from the Census of India. Safety from crime and traffic were captured as crime rates and pedestrian accident rates, respectively, acquired from State Crime Records Bureau. Greenness, built-up density, and land slope were gathered from open-source satellite imagery repository. Road intersection density was derived from OpenStreetMap. Processing and analysis differed for each dataset depending on its source and nature. Each variable showed a distinct pattern across the state. Population and residential density were found to be closely related to each other across both districts and subdistricts. They were both positively related to crime rates, pedestrian accident rates, built-up density, and intersection density, whereas negatively related to land slope and greenness across the subdistricts. Delineating the distribution of built environment variables using available and open-source data in resource-poor settings is a first in public health research among LMICs. Cost-effectiveness and reproducible nature of open-source solutions could equip researchers in resource-poor settings to identify built environment characteristics and patterns across regions.

Monitoring Islamic Archaeological Landscapes in Ethiopia Using Open Source Satellite Imagery

ABSTRACTThe African landscape is set to change dramatically in the coming years, and will have a detrimental impact on the inherent archaeological and cultural heritage elements if not monitored adequately. This paper explores how satellite imagery, in particular open source imagery (Google Earth, multispectral satellite imagery from Landsat and Sentinel-2), can be utilized to monitor and protect sites that are already known with particular reference to Islamic archaeological sites in Ethiopia. The four sites used are in different geographic and geomorphological areas: three on the Somali Plateau (Harlaa, Harar, and Sheikh Hussein), and one on the edge of the Afar Depression (Nora), and have varied histories. The results indicate that open source satellite imagery offers a mechanism for evaluating site status and conservation over time at a large scale, and can be used on data from other areas of Africa by heritage professionals in the African continent at no cost.

Appraisal of Deforestation in Murree through Open Source Satellite Imagery

Forests are the main source of income for humans and shelter for many rare in habitants. Forests act as carbon sink and help to reduce the effect of natural hazards like floods. The timber is among the main products of forests used for precious furniture. Forest must cover one third of the part of any country but in Pakistan, it covers less than 4%. If the deforestation continues with the same ratio, there will be a drastic change in the upcoming years. The main objective of the study is to determine deforestation rates in Murree for the years 1999-2015 through open source Landsat 7, 8 imageries. The areas with high deforestation rates are also delineated. Supervised classification is used to delineate the area under buildup, vegetation and water that was 83.85 km2, 760.63 km2 and 1.18 km2 respectively in 1999. Built up area increased up to 20% and forests declined by 23% in 2015 as compared to 1999. Remote sensing and GIS tools helped to investigate the deforestation rates with high accuracy.

Historic flood events in NE Romania (post-1990)

ABSTRACTUsing open-source satellite imagery like Landsat TM, ETM+ and Sentinel 2 can lead to accurate cartographic products. We mapped flood events from Siret and Prut river basins in the last 30 years based on the availability of Landsat data archive. In this area were recorded historical values in flow rates for the entire Romanian territory: 4650 m³/s on the Siret River in 2005 – the maximum value ever recorded for Romania; 4240 m³/s on the Prut in 2008 – second maximum value recorded for Romania. The most powerful floods that took place in Romania in the last years were in 1970, 1975, 1991, 2005, 2008, 2010 and 2011. In this study four years are distinguished by particularly characteristics: 1991, 2005, 2008 and 2010. Developing geo-hydrological hazard maps and adequate analysis at an appropriate scale and as quickly as possible is extremely important from an economic and social point of view.

Open-source satellite enumeration to map households: planning and targeting indoor residual spraying for malaria.

BackgroundDefining the number and location of sprayable structures (houses) is foundational to plan and monitor indoor residual spray (IRS) implementation, a primary intervention used to control the transmission of malaria. Only by mapping the location and type of all sprayable structures can IRS operations be planned, estimates of spray coverage determined, and targeted delivery of IRS to specific locations be achieved. Previously, field-based enumeration has been used to guide IRS campaigns, however, this approach is costly, time-consuming and difficult to scale. As a result, field-based enumeration typically fails to map all structures in a given area, making estimations less reliable and reducing the enumerated coverage.MethodsUsing open source satellite imagery and Geographic Information System software, satellite enumeration was conducted to guide IRS operations in 15 districts (91,302 km2) in northern Zambia during the 2014 spray season. Cost of satellite enumeration was compared to standard enumeration. Enumerated households were sampled to estimate sprayable surface area and wall type from the satellite enumeration using linear and logistic regression, respectively.ResultsIn comparison to the traditional field-based enumeration procedure, satellite-based enumeration was 22 times faster, and 10 times less costly. An estimated 98 % of the satellite enumerated buildings correctly classified roof type. Predicted surface area of each household correlated at a value of 0.91 with measured surface area of each household.ConclusionFor IRS campaigns, high quality and high coverage enumeration data aid in planning, through informed insecticide procurement. Through the identification of geographical areas and populations to target, enumeration data guide operations and assist monitoring and evaluation of IRS through the unbiased estimation of coverage achieved. Satellite enumeration represents a quick, cheap and accurate system to provide these data, and has potential applications beyond IRS for delivery of other targeted or non-targeted interventions (e.g. net distributions, mass drug administration, immunization campaigns, or even sampling frames for field studies).

Termite mounds as dominant land forms in semiarid northeastern Brazil

Open source satellite imagery has allowed the visualization and mapping of remarkable numbers of large, densely-packed termite mounds covering extensive areas in the semiarid region of northeastern Brazil, revealing tens of millions of regularly spaced termitaria (between 2 and 4 m tall, at densities of up to 35/ha). The origin, distribution, and natural history of these mounds are discussed, as well as questions regarding their age and dense packing, their sharing or appropriation by other insects, and why they attained such huge volumes in a dry landscape.