Hotspotter: A Generalizable Pipeline for Automated Detection of Subtle Volcanic Thermal Features in Satellite Images

HighlightsVegetation indices (NDVI, GNDVI, and SAVI) extracted from high-resolution satellite imagery were significantly associated with vegetation indices extracted from UAV imagery.High-resolution satellite data can be used to predict maize yield at breeding plot scale.Breeding plot sizes and the variability between maize genotypes may be associated with prediction accuracies.Abstract. The recent availability of high spatial and temporal resolution satellite imagery has widened its applications in agriculture. Plant breeding and genetics programs are currently adopting unmanned aerial vehicle (UAV) based imagery data as a complement to ground data collection. With breeding trials across multiple geographic locations, UAV imaging is not always convenient. Hence, we anticipate that, similar to UAV imaging, phenotyping of individual test plots from high-resolution satellite imagery may also provide value to plant genetics and breeding programs. In this study, high spatial resolution satellite imagery (~38 to 48 cm pixel-1) was compared to imagery acquired using a UAV for its ability to phenotype maize grown in two-row and six-row breeding plots. Statistics (mean, median, sum) of color (red, green, blue), near-infrared, and vegetation indices such as normalized difference vegetation index (NDVI), green normalized difference vegetation index (GNDVI), and soil adjusted vegetation index (SAVI) were extracted from imagery from both sources (UAV and satellite) for comparison at three time points. In general, a strong correlation between satellite and UAV imagery extracted NDVI, GNDVI, and SAVI features (especially with mean and median statistics, p < 0.001) was observed at different time points. The correlation of both UAV and satellite image features with yield potential was maximum (p < 0.001) at the third time point (milk/dough growth stages). For example, Pearson’s correlation coefficients between mean NDVI, GNDVI, and SAVI features with yield potential were 0.52, 0.54, and 0.51 for data derived from UAV imagery, and 0.34, 0.41, and 0.40 for data derived from satellite imagery, respectively. Machine learning algorithms, including least absolute shrinkage and selection operator (Lasso) regression, were evaluated for yield prediction using vegetation index features that were significantly correlated with observed yield. The relationship between satellite imagery with crop performance can be a function of plot size in addition to crop variability. Nevertheless, with the ongoing improvement of satellite technologies, there is a possibility for the integration of satellite data into breeding programs, thus improving phenotyping efficiencies. Keywords: Image processing, Machine learning, Plant breeding, Statistical analysis, Unmanned aerial vehicles.

In this research was developed the methodology useful for the future Caspian Sea coastal management what is nowadays inevitable for Azerbaijan to support decision makers in their everyday activities related with the sustainable development of coastal zones. As the material for the research were selected the reliable data sources as high-resolution satellite imagery which can be compared with the aerial photography which was always suitable for the purposes of the coastal management. The high-resolution satellite imagery satisfied the needs of coastal environmental management giving almost the same results as aerial photography. The satellite images were obtained with the coverage of pilot territories of Gizil-Agach strict nature reserve and sanctuary located along Caspian Sea coastal line. The main objective of this research was to present the up-to-date developed method of how to predict the Caspian Sea level fluctuation impacts on the coastal areas by monitoring visually results of geographical information system processing analysis based on the accurate spatial data acquired from high-resolution stereo satellite imagery. Besides nowadays acquisition of satellite imagery is easily accessible periodically what gives possibility for the permanent tracking of coastal line change dynamics. Using the advantages of high-resolution satellite imagery it is possible to extract majority of all relief topographical details along the coastal areas and over whole pilot territory using stereo plotting techniques what is very important for the further GIS modeling of Caspian Sea possible level fluctuation impacts to coastal areas. In comparison with the already implemented works in this field, the main advantage of this research is the creation of the high- accurate digital terrain model what gives us possibility for the interpolation through the defined interval and based on this principle modeling of the potential water flood and drop areas on the different levels of Caspian Sea. The results of this research presented that the processing of high-resolution stereo satellite imagery and preparation of precise stereo model, digital terrain models and digital orthophoto maps and further GIS processing of these accurate material sources are absolutely suitable methodology for the environmental modeling of potential Caspian Sea water level fluctuations. It means that it is possible to transform this photogrammetric methodology and GIS techniques into the integrated technological form that can be used by the decision makers and all other parties involved in the activities related with the Caspian Sea coastal management.

Abstract. Multiple landslide events occur often across the world which have the potential to cause significant harm to both human life and property. Although a substantial amount of research has been conducted to address mapping of landslides using Earth observation (EO) data, several gaps and uncertainties remain with developing models to be operational at the global scale. The lack of a high-resolution globally distributed and event-diverse dataset for landslide segmentation poses a challenge in developing machine learning models that can accurately and robustly detect landslides in various regions, as the limited representation of landslide and background classes can result in poor generalization performance of the models. To address this issue, we present the High-Resolution Global landslide Detector Database (HR-GLDD), a high-resolution (HR) satellite dataset (PlanetScope, 3 m pixel resolution) for landslide mapping composed of landslide instances from 10 different physiographical regions globally in South and South-East Asia, East Asia, South America, and Central America. The dataset contains five rainfall-triggered and five earthquake-triggered multiple landslide events that occurred in varying geomorphological and topographical regions in the form of standardized image patches containing four PlanetScope image bands (red, green, blue, and NIR) and a binary mask for landslide detection. The HR-GLDD can be accessed through this link: https://doi.org/10.5281/zenodo.7189381 (Meena et al., 2022a, c). HR-GLDD is one of the first datasets for landslide detection generated by high-resolution satellite imagery which can be useful for applications in artificial intelligence for landslide segmentation and detection studies. Five state-of-the-art deep learning models were used to test the transferability and robustness of the HR-GLDD. Moreover, three recent landslide events were used for testing the performance and usability of the dataset to comment on the detection of newly occurring significant landslide events. The deep learning models showed similar results when testing the HR-GLDD at individual test sites, thereby indicating the robustness of the dataset for such purposes. The HR-GLDD is open access and it has the potential to calibrate and develop models to produce reliable inventories using high-resolution satellite imagery after the occurrence of new significant landslide events. The HR-GLDD will be updated regularly by integrating data from new landslide events.

Volcanic eruptions pose a major threat to at least 800 million people. Studies revealed that ~50% of the ~1400 potentially active subaerial volcanoes still lack conventional ground-based monitoring networks. In this context, satellite data proves to be a cost-effective, yet reliable, information source for detecting early signs of volcanic activity and monitoring the evolution of eruptive events.Within the past two decades, several moderate resolution (~1 km) Mid-InfraRed (MIR) satellite-based volcano monitoring systems have been developed, mostly targeting high-temperature anomalies associated with eruptive activity. Subtle thermal anomalies, however, might occur from years to days prior major volcanic unrests and/or eruptions, and persist for a long time during the cooling stage of the erupted deposits.Studies revealed that Thermal InfraRed (TIR) bands, often characterised by higher spatial resolution (< 100 m) but lower revisit time (> 6 days), are well suited to detect subtle thermal anomalies. Yet, even in a high-temperature domain, TIR observations typically prove more effective in accurately determining the dimensions of active and cooling lava flows. Besides, high resolution TIR channels allow the retrieval of more detailed spatial information but with a temporal resolution inadequate for daily monitoring.Forefront TIR-equipped platforms, however, like the Visible Infrared Imaging Radiometer Suite (VIIRS), offer an unprecedented trade-off between spatial (375 m) and temporal resolution (up to 4 acquisitions of the same target per day), having the potential to provide accurate heat flux measurements before, during and after an eruption. Here we present a single-band TIR-based algorithm capable of detecting thermal anomalies in a broad range of volcanic settings, from crater lakes and localised low-temperature hydrothermal systems to high-temperature effusive events. The algorithm – based on temporal and contextual analyses to identify thermally anomalous pixels – can detect thermal anomalies for pixel-integrated temperatures as low as 0.5 K above the surrounding hot-spot-free background and as far as 25 km from the volcano’s summit while maintaining a false positive rate of ~2%.Results emerging from selected case studies envisage that the system will prove instrumental for detecting early signs of volcanic activity and for monitoring the evolution of thermal emissions, from unrest to eruption. Furthermore, the compilation of statistically robust multidecadal thermal datasets will provide novel insights and new perspectives into volcano monitoring, laying the ground for forthcoming higher-resolution TIR missions.

Axial Seamount, an extensively instrumented submarine volcano, lies at the intersection of the Cobb–Eickelberg hot spot and the Juan de Fuca ridge. Since late 2014, the Ocean Observatories Initiative (OOI) has operated a seven-station cabled ocean bottom seismometer (OBS) array that captured Axial’s last eruption in April 2015. This network streams data in real-time, facilitating seismic monitoring and analysis for volcanic unrest detection and eruption forecasting. In this study, we introduce a machine learning (ML)-based real-time seismic monitoring framework for Axial Seamount. Combining both supervised and unsupervised ML and double-difference techniques, we constructed a comprehensive, high-resolution earthquake catalog while effectively discriminating between various seismic and acoustic events. These events include earthquakes generated by different physical processes, acoustic signals of lava–water interaction, and oceanic sources such as whale calls. We first built a labeled ML-based earthquake catalog that extends from November 2014 to the end of 2021 and then implemented real-time monitoring and seismic analysis starting in 2022. With the rapid determination of high-resolution earthquake locations and the capability to track potential precursory signals and coeruption indicators of magma outflow, this system may improve eruption forecasting by providing short-term constraints on Axial’s next eruption. Furthermore, our work demonstrates an effective application that integrates unsupervised learning for signal discrimination in real-time operation, which could be adapted to other regions for volcanic unrest detection and enhanced eruption forecasting.

This damage assessment relied heavily on the use of remote sensing technology. Never before has the availability of high-resolution satellite and aerial imagery been so open and accessible. Data from different missions (World Bank-ImageCat-RIT Remote Sensing Mission (15cm optical and 2 pt/m2 LiDAR), Google (15cm optical), NOAA (25cm optical), Pictometry, as well as satellite imagery from GeoEye and Digitalglobe) has allowed damage from the Haiti earthquake to be viewed through multiple sensors and at different times. These multi-dimensional perspectives have been invaluable in understanding the magnitude and scope of damage caused by this earthquake.

Accurately observing and interpreting volcanic unrest phenomena contributes towards better forecasting of volcanic eruptions, thus potentially saving lives. Volcanic unrest is recorded by volcano observatories and may include seismic, geodetic, degassing and/or geothermal phenomena. The multivariate datasets are often complex and can contain a large amount of data in a variety of formats. Low levels of unrest are frequently recorded, causing the distinction between background activity and unrest to be blurred, despite the widespread usage of these terms in unrest literature (including probabilistic eruption-forecasting models) and in Volcanic Alert Level (VAL) systems. Frequencies and intensities of unrest episodes are not easily comparable over time or between volcanoes. Complex unrest information is difficult to communicate simply to civil defence personnel and other non-scientists. The Volcanic Unrest Index (VUI) is presented here to address these issues. The purpose of the VUI is to provide a semi-quantitative rating of unrest intensity relative to each volcano’s past level of unrest and to that of analogous volcanoes. The VUI is calculated using a worksheet of observed phenomena. Ranges for each phenomenon within the worksheet can be customised for individual volcanoes, as demonstrated in the companion paper for Taupo Volcanic Centre, New Zealand (Potter et al. 2015). The VUI can be determined retrospectively for historical episodes of unrest based on qualitative observations, as well as for recent episodes with state-of-the-art monitoring. This enables a long time series of unrest occurrence and intensity to be constructed and easily communicated to end users. The VUI can also assist with VAL decision-making. We present and discuss two approaches to the concept of unrest.

<p>Volcanic eruptions present serious risk to human life and infrastructure. This risk can be minimized by improving eruption forecasts, which in turn requires increasing our capabilities to detect volcanic unrest and a better understanding of the physicochemical processes governing magma-hydrothermal interactions. The improvement of eruption forecasting techniques is especially important as some volcanic eruptions can occur with little to no precursory warning signs. That was the case of the most recent eruption at Okmok caldera, which took place in 2008 between July 12 – August 23, with a volcanic explosivity index of 4. This eruption highlighted the need to develop new methods to detect precursory activity and unrest.</p><p>Recently, through the analysis of satellite-based thermal spectroscopy data from MODIS instruments, Girona <em>et al.</em> (2021) found that low-temperature thermal anomalies along the flanks of volcanoes can predate their eruptions. In this work, we use an updated version of the method presented in Girona <em>et al.</em> (2021) to analyze the spatiotemporal distribution of low-temperature thermal anomalies at Okmok Caldera between July of 2002 and November of 2021. Preliminary analysis shows ~1-1.3 degrees of warming at Cone A in the ~3 years leading up to the 2008 eruption. This analysis also shows a warming trend in the caldera at several cones (D, E, A, and Ahmanilix), peaking in 2014, with brightness temperatures increasing by ~1-1.4 degrees for ~2 years (correlating with an observed inflation event); along with current warming at the same cones of ~0.8-1.2 degrees beginning in ~2017.</p><p>We propose that the low-temperature thermal anomalies observed at different cones of Okmok caldera are linked to the latent heat released during the condensation of magmatic and/or hydrothermal water vapor in the subsurface. In particular, we design a 1-dimensional thermal diffusion model to quantify how long it will take for the surface ground temperature to increase by one kelvin in response to the subsurface condensation of water vapor. Our preliminary analysis shows that, for realistic values of the parameters involved, the surface requires ~3.3 years to increase its temperature by one kelvin in response to a diffuse H2O flux of 161.5 kg/s condensing at 30m depth, and ~21.7 years for the surface to increase by one kelvin in response to the same gas flux condensing at 60m depth. The observed low-temperature thermal anomalies at Okmok are therefore consistent with the condensation of magmatic and/or hydrothermal water vapor at no more than a few tens of meters depth below the surface.</p><p>This work provides further insight into how volcanic hydrothermal subsurface processes manifest as thermal anomalies on the surface, and how these thermal anomalies can be used to detect unrest at Okmok and other active volcanoes. In the future, we aim to integrate the spatiotemporal distribution of low-temperature thermal anomalies with deformation, seismic signals, and diffuse gas emissions prior to and during eruptions.</p><p> </p><p>Girona, T., Realmuto, V. & Lundgren, P. Large-scale thermal unrest of volcanoes for years prior to eruption. <em>Nat. Geosci.</em> <strong>14, </strong>238–241 (2021). https://doi.org/10.1038/s41561-021-00705-4.</p>

Volcanic eruptions are often preceded by small changes in the shape of the volcano. Such volcanic deformation may be measured using precise surveying techniques and analysed to better understand volcanic processes. Complicating the matter is the fact that deformation events (e.g., inflation or deflation) may result from magmatic, non-magmatic or mixed/hybrid sources. Using spatial and temporal patterns in volcanic deformation data and mathematical models it is possible to infer the location and strength of the subsurface driving mechanism. This can provide essential information to inform hazard assessment, risk mitigation and eruption forecasting. However, most generic models over-simplify their representation of the crustal conditions in which the deformation source resides. We present work from a selection of studies that employ advanced numerical models to interpret deformation and gravity data. These incorporate crustal heterogeneity, topography, viscoelastic rheology and the influence of temperature, to constrain unrest source parameters at Uturuncu (Bolivia), Cotopaxi (Ecuador), Soufriere Hills (Montserrat), and Teide (Tenerife) volcanoes. Such model complexities are justified by geophysical, geological, and petrological constraints. Results highlight how more realistic crustal mechanical conditions alter the way stress and strain are partitioned in the subsurface. This impacts inferred source locations and magmatic pressures, and demonstrates how generic models may produce misleading interpretations due to their simplified assumptions. Further model results are used to infer quantitative and qualitative estimates of magma supply rate and mechanism, respectively. The simultaneous inclusion of gravity data alongside deformation measurements may additionally allow the magmatic or non-magmatic nature of the source to be characterised. Together, these results highlight how models with more realistic, and geophysically consistent, components can improve our understanding of the mechanical processes affecting volcanic unrest and geodetic eruption precursors, to aid eruption forecasting, hazard assessment and risk mitigation.

Identification of lake trout Salvelinus namaycush spawning habitat in northern Lake Huron using high-resolution satellite imagery

Deep learning for multi-modal classification of cloud, shadow and land cover scenes in PlanetScope and Sentinel-2 imagery

Due to its high spatial resolution, broad spatial coverage, and cost-effectiveness, commercial satellite imagery is rapidly becoming a key component of biological monitoring in the Antarctic. While considerable success in surveying emperor penguins (Aptenodytes forsteri) has been facilitated by their large size and the visual simplicity of their habitat, there has been considerably less progress in mapping colonies on the Antarctic Peninsula and associated sub-Antarctic islands where smaller penguin species breed on topographically complex terrain composed of mixed substrates. Here, we demonstrate that Adelie penguin (Pygoscelis adeliae), chinstrap penguin (P. antarcticus), gentoo penguin (P. papua), and macaroni penguin (Eudyptes chrysolophus) colonies can be detected by high-resolution (2-m multispectral, 40–50-cm panchromatic) satellite imagery and that under ideal conditions, such imagery is capable of distinguishing among groups of species where they breed contiguously. To demonstrate the potential for satellite imagery to estimate penguin population abundance, we use satellite imagery of Paulet Island (63°35′S, 55°47′W) to estimate a site-wide population of 115,673 (99,222–127,203) breeding pairs of Adelie penguins.

This paper presents an automatic procedure for rapid building extraction from optical very high resolution (VHR) satellite imagery. Classical extraction models are always complex and time-consuming. The optimized process of building extraction consists of three main rapid stages: edge-preserving and smoothing bilateral filter, line segment detection, perceptual grouping polygonal building boundary. Firstly, we use bilateral filter to smooth original image with edge-preserving. Secondly, a state-of-the-art line segment detector (LSD) algorithm gives highly accurate building contour segments. Finally, we apply the perceptual grouping approach based on graph search to organize detected contour line segments of interested buildings. We test our method on optical VHR QuickBird satellite imagery and obtain promising experimental results with overall accuracy of 79.1%, which confirm the effectiveness and robustness of this linear-time procedure.

Large-scale monitoring of coppice forest clearcuts by multitemporal very high resolution satellite imagery. A case study from central Italy

Mapping and characterizing Arctic beaded streams through high resolution satellite imagery