ESSI1.12

An breadboard for end-to-end (E2E) Marine Litter Optical Performance Simulations (ML-OPSI) is being designed in the frame of the ESA Open Space Innovation Platform (OSIP) Campaign to support Earth Observation (EO) scientists with the design of computational experiments for Operations Research. The ML-OPSI breadboard will estimate Marine Litter signal at Top-Of-Atmosphere (TOA) from a set of Bottom-Of-Atmosphere (BOA) scenarios representing the various case studies by the community (e.g., windrows, frontal areas, river mouths, sub-tropical gyres), coming from synthetic data (computer-simulated) or from real observations. It is a modular, pluggable and extensible framework, promoting re-use and be adapted for different missions, sensors and scenarios.

The breadboard consists of (a) the OPSI components for the simulation i.e. the process of using a model to study the characteristics of the system by manipulating variables and by studying the properties of the model allowing an evaluation to optimise performance and make predictions about the real system; and (b) the Marine Litter model components for the detection of marine litter. It shall consider the changes caused in the water reflectance and properties due to marine litter, exploiting gathered information of plastic polymers, different viewing geometries, and atmospheric conditions as naturally occurring. The modules of the breadboard include a Scenario Builder Module (SB) with maximum spatial resolution and best modelling as possible of the relevant physical properties, which for spectral sensors could include high spatial resolution and high spectral density/resolution BOA radiance simulations in the optical to SWIR bands; a Radiative Transfer Module (RTM) transforming water-leaving to TOA reflectance for varying atmospheric conditions and observational geometries; a Scene Generator Module (SGM) which could use Sentinel-2, Landsat, or PRISMA data as reference or any other instrument as pertinent; a Performance Assessment Module (PAM) for ML detection that takes into account the variability of the atmosphere, the sunlight & skylight at BOA, the sea-surface roughness with trains of wind waves & swells, sea-spray (whitecaps), air bubbles in the mixed layer, marine litter dynamics as well as instrumental noise to assess marine litter detection feasibility.

Marine Litter scenarios of reference shall be built based on in-situ campaigns, to reflect the true littering conditions at each case, both in spatial distribution and composition. The breadboard shall be validated over artificial targets at sea in field campaigns as relevant. This might include spectral measurements from ASD, on-field radiometers, and cameras on UAVs, concomitant with Copernicus Sentinel-2 acquisitions. Combined, they can be used to estimate atmospheric contribution and assess performance of the testes processing chain.

This activity collaborates on the ““Remote Sensing of Marine Litter and Debris” IOCCG taskforce.

How to cite: Emsley, S., Arias, M., Papadopoulou, T., and Martin-Lauzer, F.-R.: Simulating Marine Litter observations from space to support Operations Research, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-15661, https://doi.org/10.5194/egusphere-egu21-15661, 2021.

Automatically extracting buildings from remote sensing images (RSI) plays important roles in urban planning, population estimation, disaster emergency response, etc. With the development of deep learning technology, convolutional neural networks (CNN) with better performance than traditional methods have been widely used in extracting buildings from remote sensing imagery (RSI). But it still faces some problems. First of all, low-level features extracted by shallow layers and abstract features extracted by deep layers of the artificial neural network could not be fully fused. it makes building extraction is often inaccurate, especially for buildings with complex structures, irregular shapes and small sizes. Secondly, there are so many parameters that need to be trained in a network, which occupies a lot of computing resources and consumes a lot of time in the training process. By analyzing the structure of the CNN, we found that abstract features extracted by deep layers with low geospatial resolution contain more semantic information. These abstract features are conducive to determine the category of pixels while not sensitive to the boundaries of the buildings. We found the stride of the convolution kernel and pooling operation reduced the geospatial resolution of feature maps, so, this paper proposed a simple and effective strategy—reduce the stride of convolution kernel contains in one of the layers and reduced the number of convolutional kernels to alleviate the above two bottlenecks. This strategy was used to deeplabv3+net and the experimental results for both the WHU Building Dataset and Massachusetts Building Dataset. Compared with the original deeplabv3+net the result showed that this strategy has a better performance. In terms of WHU building data set, the Intersection over Union (IoU) increased by 1.4% and F1 score increased by 0.9%; in terms of Massachusetts Building Dataset, IoU increased by 3.31% and F1 score increased by 2.3%.

How to cite: Chen, M., Wu, J., and Tian, F.: Reducing the stride of the convolution kernel: a simple and effective strategy to increase the performance of CNN in building extraction from remote sensing image, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-10783, https://doi.org/10.5194/egusphere-egu21-10783, 2021.

Long time series of essential climate variables (ECVs) derived from satellite data are key to climate research. SemantiX is a research project to establish, complement and expand Advanced Very High Resolution Radiometer (AVHRR) time series using Copernicus Sentinel-3 A/B imagery, making them and derived ECVs accessible using a semantic Earth observation (EO) data cube. The Remote Sensing Research Group at the University of Bern has one of the longest European times series of AVHRR imagery (1981-now). Data cube technologies are a game changer for how EO imagery are stored, accessed, and processed. They also establish reproducible analytical environments for queries and information production and are able to better represent multi-dimensional systems. A semantic EO data cube is a newly coined concept by researchers at the University of Salzburg referring to a spatio-temporal data cube containing EO data, where for each observation at least one nominal (i.e., categorical) interpretation is available and can be queried in the same instance (Augustin et al. 2019). Offering analysis ready data (i.e., calibrated and orthorectified AVHRR Level 1c data) in a data cube along with semantic enrichment reduces barriers to conducting spatial analysis through time based on user-defined AOIs.

This contribution presents a semantic EO data cube containing selected ECV time series (i.e., snow cover extent, lake surface water temperature, vegetation dynamics) derived from AVHRR imagery (1981-2019), a temporal and spatial subset of AVHRR Level 1c imagery (updated after Hüsler et al. 2011) from 2016 until 2019, and, for the later, semantic enrichment derived using the Satellite Image Automatic Mapper (SIAM). SIAM applies a fully automated, spectral rule-based routine based on a physical-model to assign spectral profiles to colour names with known semantic associations; no user parameters are required, and the result is application-independent (Baraldi et al. 2010). Existing probabilistic cloud masks (Musial et al. 2014) generated by the Remote Sensing Research Group at the University of Bern are also included as additional data-derived information to support spatio-temporal semantic queries. This implementation is a foundational step towards the overall objective of combining climate-relevant AVHRR time series with Sentinel-3 imagery for the Austrian-Swiss alpine region, a European region that is currently experiencing serious changes due to climate change that will continue to create challenges well into the future.

Going forward, this semantic EO data cube will be linked to a mobile citizen science smartphone application. For the first time, scientists in disciplines unrelated to remote sensing, students, as well as interested members of the public will have direct and location-based access to these long EO data time series and derived information. SemantiX runs from August 2020-2022 funded by the Austrian Research Promotion Agency (FFG) under the Austrian Space Applications Programme (ASAP 16) (project #878939) in collaboration with the Swiss Space Office (SSO).

How to cite: Augustin, H., Sudmanns, M., Weber, H., Baraldi, A., Wunderle, S., Neuhaus, C., Reichel, S., van der Meer, L., Hummer, P., and Tiede, D.: SemantiX: a cross-sensor semantic EO data cube to open and leverage AVHRR time-series and essential climate variables with scientists and the public, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-12722, https://doi.org/10.5194/egusphere-egu21-12722, 2021.

Nitrogen dioxide (NO²) is one of the main air quality pollutants of concern in many urban and industrial areas worldwide. Being emitted by fossil fuel burning activities including mainly road traffic, the NO² pollution is responsible for population health degradation and secondary pollutants formation as nitric acid and ozone. In the European region, almost 20 countries exceeded in 2017 the NO² annual limit values imposed by European Commission Directive 2008/50/EC (EEA, 2019). Therefore, NO² pollution monitoring and regulation is a necessary task to help decision makers to search for a sustainable solution for environmental quality and population health status improvement. In this study, we propose a comparative analysis of the tropospheric NO² column density spatial configuration over Europe between similar periods from 2019 and 2020, based on ESA Copernicus Sentinel-5P products. Our results highlight the NO² pollution dynamics over the abrupt transition from a normal condition situation to the COVID-19 outbreak context, characterized by short-time decrease of traffic intensities and industrial activities, this situation being also reflected by the national level statistics referring to COVID-19 cases and economic indicatiors. The validation approach provides high correlation between TROPOMI derived data and independent data from ground-based observations with encouraging values of the R² ranging between 0.5 and 0.75 in different locations.  

How to cite: Vîrghileanu, M., Săvulescu, I., Mihai, B.-A., Nistor, C., and Dobre, R.: Using Sentinel-5P time-series products for Nitrogen Dioxide (NO2) Spatio-Temporal Analysis over Europe During the Coronavirus Pandemic Lockdown, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-10882, https://doi.org/10.5194/egusphere-egu21-10882, 2021.

The climate is strongly affected by interaction with clouds. To reduce major errors in climate predictions, this interaction requires a much finer understanding of cloud physics than current knowledge. Current knowledge is based on empirical remote sensing data that is analyzed under the assumption that the atmosphere and clouds are made of very broad and uniform layers. To help to overcome this problem, 3D scattering computed tomography (CT) has been suggested as a way to study clouds. 

CT is a powerful way to recover the inner structure of three dimensional (3D) volumetric heterogeneous objects. CT has extensive use in many research and operational domains. Aside from its common usage in medicine, CT is used for sensing geophysical terrestrial structures, atmospheric pollution and fluid dynamics. CT requires imaging from multiple directions and in nearly all CT approaches, the object is considered static during image acquisition. However, in many cases, the object changes while multi-view images are acquired sequentially. Thus, an effort has been invested to expand 3D CT to four-dimensional (4D) spatiotemporal CT. This effort has been directed at linear CT modalities. Since linear CT is computationally easier to handle, it has been a popular method for medical imaging. However, these linear CT modalities do not apply to clouds: clouds constitute a scattering medium, and therefore radiative transfer is non-linear in the clouds’ content.

This work focuses on the challenge of 4D scattering CT of clouds. Scattering CT of clouds requires high-resolution multi-view images from space. There are spaceborne and high-altitude systems that may provide such data, for example AirMSPI, MAIA, HARP and AirHARP. An additional planned system is the CloudCT formation, funded by the ERC. However, these systems are costly. Deploying them in large numbers to simultaneously acquire images of the same clouds from many angles can be impractical. Therefore, the platforms are planned to move above the clouds: a sequence of images is taken, in order to span and sample a wide angular breadth. However, the clouds evolve while the angular span is sampled.

We pose conditions under which this task can be performed. These regard temporal sampling and angular breadth, in relation to the correlation time of the evolving cloud. Then, we generalize scattering CT. The generalization seeks spatiotemporal recovery of the cloud extinction field in high resolution (10m), using data taken by a small number of moving cameras. We present an optimization-based method to reach this, and then demonstrate the method both in rigorous simulations and on real data.

How to cite: Ronen, R., Schechner, Y. Y., and Eytan, E.: Spatiotemporal tomography based on scattered multiangular signals and its use for resolving evolving clouds using moving platforms, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-10892, https://doi.org/10.5194/egusphere-egu21-10892, 2021.

Urbanization and the trend of people moving to cities often leads to problematic traffic conditions, which can be very challenging for traffic management. It can hamper the flow of people and goods, negatively affecting businesses through delays and the inability to estimate travel times and thus plan, as well as the environment and health of population due to increased fuel consumption and subsequent air pollution. Many cities have a policy and rules to manage traffic, ranging from standard traffic lights to more dynamic and adaptable solutions involving in-road sensors or cameras to actively modify the duration of traffic lights, or even more sophisticated IoT solutions to monitor and manage the conditions on a city-wide scale. The core to these technologies and to decision making processes is the availability of reliable data on traffic conditions, and better yet real-time data. Thus, a lot of cities are still coping with the lack of good spatial and temporal data coverage, as many of these solutions are requiring not only changes to the infrastructure, but also large investments.

One approach is to exploit the current and the forthcoming advancements made available by Earth Observation (EO) satellite technologies. The biggest advantage is EOs great spatial coverage ranging from a few km² to 100 km² per image on a spatial resolution down to 0.3m, thus allowing for a quick, city-spanning data collection. Furthermore, the availability of imaging sensors covering specific bands allows the constituent information within an image to be separated and the information to be leveraged.

In this respect, we present the findings of our work on multispectral image sets collected on three occasions in 2019 using very high resolution WorldView-3 satellite. We apply a combination of machine learning and PCA methods to detect vehicles and devise their kinematic properties (e.g., movement, direction, speed), only possible with satellites with a specific design allowing for short time lags between imaging in different spectral bands. As these data basically constitute a time-series, we will discuss how the results presented fully apply to the forthcoming WorldView-Legion constellation of satellites providing up to 15 revisits per day, and thus near-real time traffic monitoring and its impact on the environment.

How to cite: Duro, R., Neubauer, G., and Bojor, A.-I.: The potential of monitoring traffic conditions up to 15 times a day using sub-meter resolution EO images, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-12831, https://doi.org/10.5194/egusphere-egu21-12831, 2021.

Processing, handling and visualising the large data volume produced by satellite altimetry missions is a challenging task. A reference tool for the visualisation of satellite laser altimetry data is the OpenAltimetry platform, a tool that provides altimetry-specific data from the Ice, Cloud, and land Elevation Satellite (ICESat) and ICESat-2 satellite missions through a web-based interactive interface. However, by focusing only on altimetry data, that tool leaves out access to many other equally important information existing in the data products from both missions.

The main objective of the work reported here was the development of a new web-based tool, called ICEComb, that offers end users the ability to access all the available data from both satellite missions, visualise and interact with them on a geographic map, store the data records locally, and process and explore data in an efficient, detailed and meaningful way, thus providing an easy-to-use software environment for satellite laser altimetry data analysis and interpretation.

The proposed tool is intended to be mainly used by researchers and scientists to aid their work using ICESat and ICESat-2 data, offering users a ready-to-use system to rapidly access the raw collected data in a visually engaging way, without the need to have prior understanding of the format, structure and parameters of the data products. In addition, the architecture of the ICEComb tool was developed with possible future expansion in mind, for which well-documented and standard languages were used in its implementation. This allows, e.g., to extend its applicability to data from other satellite laser altimetry missions and integrate models that can be coupled with ICESat and ICESat-2 data, thus expanding and enriching the context of studies carried out with such data.

The use of the ICEComb tool is illustrated and demonstrated by its application to ICESat/GLAS measurements over Lake Mai-Ndombe, a large and shallow freshwater lake located within the Ngiri-Tumba-Maindombe area, one of the largest Ramsar wetlands of international importance, situated in the Cuvette Centrale region of the Congo Basin.

Keywords: Laser altimetry, ICESat/GLAS, software tool design, data visualization, Congo Basin.

Acknowledgement. This work was partially supported by the Portuguese Foundation for Science and Technology (FCT) through LARSyS − FCT Pluriannual funding 2020−2023.

How to cite: Silva, B., Guerreiro Lopes, L., and Campos, P.: ICEComb − A New Software Tool for Satellite Laser Altimetry Data Processing and Visualisation, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-13727, https://doi.org/10.5194/egusphere-egu21-13727, 2021.

The use of remote sensing in mineral detection and lithological mapping has become a generally accepted augmentative tool in exploration. With the advent of multispectral sensors (e.g. ASTER, Landsat, Sentinel and PlanetScope) having suitable wavelength coverage and bands in the Shortwave Infrared (SWIR) and Thermal Infrared (TIR) regions, multispectral sensors have become increasingly efficient at routine lithological discrimination and mineral potential mapping. It is with this paradigm in mind that this project sought to evaluate and discuss the detection and mapping of vanadium bearing magnetite, found in discordant bodies and magnetite layers, on the Eastern Limb of the Bushveld Complex. The Bushveld Complex hosts the world’s largest resource of high-grade primary vanadium in magnetitite layers, so the wide distribution of magnetite, its economic importance, and its potential as an indicator of many important geological processes warranted the delineation of magnetite.

The detection and mapping of the vanadium bearing magnetite was evaluated using specialized traditional, and advanced machine learning algorithms. Prior to this study, few studies had looked at the detection and exploration of magnetite using remote sensing, despite remote sensing tools having been regularly applied to diverse aspects of geosciences. Maximum Likelihood, Minimum Distance to Means, Artificial Neural Networks, Support Vector Machine classification algorithms were assessed for their respective ability to detect and map magnetite using the PlanetScope data in ENVI, QGIS, and Python. For each classification algorithm, a thematic landcover map was attained and the accuracy assessed using an error matrix, depicting the user's and producer's accuracies, as well as kappa statistics.

The Maximum Likelihood Classifier significantly outperformed the other techniques, achieving an overall classification accuracy of 84.58% and an overall kappa value of 0.79. Magnetite was accurately discriminated from the other thematic landcover classes with a user’s accuracy of 76.41% and a producer’s accuracy of 88.66%. The erroneous classification of some mining activity pixels as magnetite in the Maximum Likelihood was inherent to all classification algorithms. The overall results of this study illustrated that remote sensing techniques are effective instruments for geological mapping and mineral investigation, especially in iron oxide mineralization in the Eastern Limb of Bushveld Complex. 

How to cite: Twala, M., Roberts, J., and Munghemezulu, C.: Use of multispectral remote sensing data to map magnetite bodies in the Bushveld Complex, South Africa: a case study of Roossenekal, Limpopo., EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-7932, https://doi.org/10.5194/egusphere-egu21-7932, 2021.