Potential supervisory team: Stefan Lang, Dirk Tiede, Lorenz Wendt, Getachew GellaGlobally there is a profound number of the global population displaced from their homes because of natural and anthropogenic reasons. These forcibly displaced segment of the global population resides either in IDP sites or refugee camps if they cross international borders. Dwelling extraction and camp monitoring for a specific single camp can be efficiently done by using deep learning models and high-resolution satellite images and annotated labels to train the model. The biggest problem is when new IDPs/refugee camps are happening and rapid information retrieval is needed for emergency response for humanitarian response. The proposed work intends to leverage the learned skills of the model trained in IDP/refugee camps to extract features in completely unseen images taken from other new IDP/refugee camps. One of the constraints which limit this learning process is that dwelling characteristics in different refugee camps would be completely different which is governed by the medium of the material used to construct the houses, its contrast with the surrounding information, sizes of houses, and spacing of houses (See Fig. 1). A model trained in one site is said to be universally capable of extracting features if and only if it is spatially transferable and able to work very well in refugee camps that are located in other geographic locations. Therefore, the objective of the proposed work will be firstly assessing the spatial transferability of the instance segmentation model and designing implementation strategies that can improve the spatial transferability of the model.Figure 1: Dwelling type and structures in three different refugee camps

Potential supervisory team: Stefan Lang, Dirk Tiede, Lorenz Wendt, Getachew Gella

Deep learning-based dwelling extraction in internally displaced population sites (IDPs) and refugee camps, satellite images could be obtained with different bit depth, processing labels (calibration, atmospheric correction, and further enhancement). With the same token, labeled/annotated data may also be obtained from crowd sources or in-house generated databases with a different label of quality where there might be class mislabelling, imprecise dwelling boundaries (Fig 1), completely unlabelled objects, and others. Therefore, the main theme of the proposed topic will be investigating the implication of investing intensive reprocessing phases on the image side or simply using a co-registered image on the performance of dwelling extraction. Undertaking intensive image pre-processing steps and the generation of the very precise respective object labels for model training and testing demands a profound amount of time. Especially when it comes to operational settings in emergency response for humanitarian response, it could be a pressing challenge for rapid mapping and information retrieval. Therefore, the proposed study will focus on two issues. Firstly, a thorough investigation will be done on trade-offs of using a clean but small amount of labeled sample data and a large number of grey samples with less precise annotation quality for dwelling extraction using deep learning models. Second, a substantial effort would be invested in the creation of robust deep learning strategies that could leverage from large grey data in a learning phase

Figure 1: Satellite image of dwelling features(A) and image with annotated labels with shifted object boundary and two mislabeled objects

Suggested/supervised by: Yunya Gao, Stefan Lang

Due to human armed conflicts, human rights violations, persecution, or environmental degradation, numerous people are being forced to flee their homes globally. They are called refugees if they have crossed international borders, or internally displaced persons (IDPs) in the case they are on the run within their home countries. Refugee/IDP dwellings are commonly used as temporary shelters for these displaced people. Detailed information on refugee/IDP dwelling infrastructure, the population in need and their spatial distribution is important for planning humanitarian actions (Sprohnle, Fuchs, & Aravena Pelizari, 2017). However, during a crisis, such critical information is usually hard to access by fieldwork (Witmer, 2015). Therefore, detecting refugee/IDP camps through remote sensing (RS) techniques has attracted a lot of attention.

Very high spatial resolution optical (VHSRO) satellite imagery is considered as the major source of information to identify the number and types of structures (Ghorbanzadeh, Tiede, Wendt, Sudmanns, & Lang, 2021). Visual interpretation of VHSRO imagery is recognized to produce accurate refugee/IDP camp mapping results (Witmer, 2015). However, it is thought to be highly reliant on context knowledge, time-consuming and labour intensive (Witmer, 2015). Recently, deep learning (DL) methods have been proved to outperform many other automatic processing methods in many remote sensing domains. However, DL methods usually require a large amount of data and high computational cost (Zhu et al., 2017). In the past few decades, ZGIS has assisted Doctors without borders (Médecins Sans Frontières, MSF) in collecting a large amount of data of refugee/IDP camps together with VHSRO imagery for many countries. Therefore, DL methods have high potentials in detecting refugee/IDP dwellings. However, by looking through the collected data, there are usually multiple classes of dwellings. The amount of data for each class is imbalanced. After some initial experiments, it is found out that the accuracy of minority classes (e.g. Tukul) is usually much lower than majority classes (e.g. bright dwellings). It is still unknown how to deal with imbalanced data issues in refugee/IDP dwelling detection by using DL methods. Improving the accuracy of prediction results of minority classes may be valuable for planning humanitarian actions.

Figure 1. Three types of refugee/IDP dwellings represented from very high spatial resolution satellite imagery

Reference:

Ghorbanzadeh, O., Tiede, D., Wendt, L., Sudmanns, M., & Lang, S. (2021). Transferable instance segmentation of dwellings in a refugee camp - integrating CNN and OBIA. European Journal of Remote Sensing, 54(sup1), 127–140. https://doi.org/10.1080/22797254.2020.1759456 Sprohnle, K., Fuchs, E. M., & Aravena Pelizari, P. (2017). Object-Based Analysis and Fusion of Optical and SAR Satellite Data for Dwelling Detection in Refugee Camps. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 10(5), 1780–1791. https://doi.org/10.1109/JSTARS.2017.2664982 Witmer, F. D. W. (2015). Remote sensing of violent conflict: eyes from above. International Journal of Remote Sensing, 36(9), 2326–2352. https://doi.org/10.1080/01431161.2015.1035412 Zhu, X. X., Tuia, D., Mou, L., Xia, G. S., Zhang, L., Xu, F., & Fraundorfer, F. (2017). Deep learning in remote sensing: a review. ArXiv, (December). https://doi.org/10.1109/MGRS.2017.2762307

Suggested/supervised by: Linda Menk

Background:

Despite huge achievements in the prevention and control of malaria, still more than 200 million humans globally fall ill with it every year. About 95% of the cases and related deaths happen in sub-Saharan Africa. The population group at highest risk are children under the age of 5.Malaria is transmitted by certain mosquito species which are dependent on certain habitat preconditions. For example, Vegetation health (VH) was found to be a good indicator of annual mosquito activity and related spikes in malaria outbreaks.

Source: Malaria Atlas Project (https://malariaatlas.org/)

The European flagship programme on Earth Observation Copernicus offers a growing variety of global datasets which are derived from satellite imagery, such as the Vegetation Condition Index and the Vegetation Productivity Index. The departure point for this master thesis could be the exploration of the global datasets offered by Copernicus (and others) and their potential to function as malaria risk indicators. After exploring which datasets could be put to use and why, the data should be retrieved, integrated and visualized for a specific region of interest in Africa. The master thesis will be part of a malaria case study which is currently conducted together with Doctors without borders/Médecins sans Frontières (MSF), as part of the Christian Doppler Laboratory “GEOHUM”. The results will help MSF to focus their malaria prevention activities to areas which face the highest risk.

Further reading: 

https://land.copernicus.eu/global/themes/vegetation https://link.springer.com/chapter/10.1007/978-3-030-46020-4_6 https://www.who.int/teams/global-malaria-programme/reports/world-malaria-report-2020/

Suggested/supervised by: Lorenz Wendt

Background:In the northern province of Mozambique, Cabo Delgado, Islamic extremist armed groups started aggressive attacks on civilians in October 2017. These attacks have continued in 2021, causing insecurity and displacement within the area, food shortages, and further suffering by the backlash of government forces. An estimated 713,000 people have been internally displaced, mostly from the disputed coastal areas towards more inland regions according to estimates by UNHCR and IOM.

The conflict zone is partly considered a “no access /hard to reach area”, making it difficult to get a comprehensive overview of the situation for humanitarian actors like MSF or UN organisations. The task is therefore to leverage EO data to map the effect of the conflict on people, villages, transport infrastructure and agricultural production. Multitemporal optical imagery from Sentinel-2 and potentially Planetscope shall be used to identify abandoned villages and roads by mapping overgrowth. Newly formed settlements in the safer regions surrounding the conflict zone may be detected by bushland or forest clearing. A reduction of active cropland might indicate displacement of people and/or food insecurity. These analyses can be carried out in Google Earth Engine or other cloud processing platforms, following strategies employed also by the World Food Programme.

In addition, VHR images might be analyzed to map dwellings (tents, huts, etc) of IDPs in selected hotspot locations, for example Pemba City (MOZ). Given the breadth of EO and GI data available, the exact scope, dataset and methodology will be discussed with the MSc candidate individually.

Further reading:

https://www.acaps.org/country/mozambique/crisis/violent-insurgency-in-cabo-delgado https://reliefweb.int/report/mozambique/urgent-needs-mozambique-cabo-delgado-situation-7-may-2021 https://displacement.iom.int/reports/iom-dtm-baseline-assessment-report-round-11-march-2021 https://www.wfp.org/publications/wfp-mali-satellite-imagery

Suggested / supervised by: Andreas Braun

Background: In early 2021 an unusually high amount of rainfall hit the city of Casablanca with 250 mm of precipitation falling from January 6th to 11th , which accounts for about 70% of this region’s yearly average. This caused heavy flooding in the city with immense damages to infrastructures and human safety. While open water bodies can be mapped with radar data, urban areas are still challenging because of the signal from buildings which superimposes and obscures the signal from the flooded ground. An approach based on interferometric coherence is to be tested for its suitability to detect water bodies in urban areas based on three COSMO-SkyMed images (two before and one during the flood event). Coherence images of the pair without flooding is compared with the pair with flooding and analyzed for indications of inner-urban flooding. The results are validated against references from the DisasterCharter, photographs and media reports.

Suggested reading:

Pulvirenti, L., Chini, M., Pierdicca, N., & Boni, G. (2015). Use of SAR data for detecting floodwater in urban and agricultural areas: The role of the interferometric coherence. IEEE Transactions on Geoscience and Remote Sensing, 54(3), 1532-1544. Li, Y., Martinis, S., & Wieland, M. (2019). Urban flood mapping with an active self-learning convolutional neural network based on TerraSAR-X intensity and interferometric coherence. ISPRS Journal of Photogrammetry and Remote Sensing, 152, 178-191. 

Start: anytime

Prerequisites/qualification: Basic knowledge and interest in radar remote sensing and VHR image analysis.

Suggested / supervised by: Andreas Braun

Background: While VHR optical data are already widely employed for the mapping of informal settlements, makeshift buildings and refugee camps, the role of radar imagery is still rather unexplored. Satellites such as TerraSAR-X allow the acquisition of images at sub-meter resolutions, but the interpretation and analysis is challenging because of the side-looking geometry of the radar and backscatter inambiguities between buildings and vegetation. The thesis should develop a routine for the mapping of buildings from VHR SAR images of Dagahaley (Kenya), Minawao (Cameroon) or Kutupalong (Bangladesh). There are multiple images per camp, partly acquired from different angles at different polarizations. Additional layers can be computed from image textures, transformations or interferometric coherence. Building footprints are also available as reference data. The identification can be based on either OBIA, pattern recognition or convolutional neural networks (CNNs).

Suggested reading:

Braun, A. (2021). Extraction of Dwellings of Displaced Persons from VHR Radar Imagery – A Review on Current Challenges and Future Perspectives. GI_Forum 9(1), 201-208. 

Start: anytime

Prerequisites/qualification: Basic knowledge and interest in radar remote sensing and VHR image analysis.

Suggested / supervised by: Andreas Braun

Background: Several studies investigate the changes in agricultural patterns resulting from armed conflicts (see references) based on optical satellite information. The study shall investigate the potential of ALOS-2 imagery for the mapping and change detection of agricultural production in central Syria during the civil war and relate the changes to different phases of occupation by the Islamic State. For this, 20 images were acquired in StripMap mode (HH/VH polarization, 10 m spatial resolution) between 2014 and 2020. The overall goal is the development of a routine for multi-temporal change detection and interpretation, based on ALOS-2 data, optionally in combination with Sentinel-1 acquisitions (VV/VH) of similar acquisition date as demonstrated by Braun et al. (2019).

Suggested reading: Braun, A., Fakhri, F., & Hochschild, V. (2019). Refugee camp monitoring and environmental change assessment of Kutupalong, Bangladesh, based on radar imagery of sentinel-1 and ALOS-2. Remote Sensing, 11(17), 2047.  Jaafar, H. H., & Woertz, E. (2016). Agriculture as a funding source of ISIS: A GIS and remote sensing analysis. Food Policy, 64, 14-25. Eklund, L., Degerald, M., Brandt, M., Prishchepov, A. V., & Pilesjö, P. (2017). How conflict affects land use: agricultural activity in areas seized by the Islamic State. Environmental Research Letters, 12(5), 054004. 

Start: anytime

Prerequisites/qualification: Basic knowledge and interest in radar remote sensing and multi-temporal analyses

Suggested / supervised by: Stefan Lang, Dirk Tiede, Yunya Gao

Background: The Christian Doppler Laboratory on Earth Observation for Humanitarian Action – GEOHUM – seeks to advance technologies at the interface of EO*GI and AI (artificial intelligence). The applications range from mission planning and operations in crisis intervention to population estimation for food distribution or vaccination campaigns. Our vision is to enhance technical and organisational capacities matching specific needs from humanitarian organisations, in particular our partner MSF (Doctors Without Borders). Envisaged outcomes are a fundamental scientific substantiation and the development and innovative use of relevant information products to optimize aid delivery in conflict and humanitarian disaster situations.

With promising results obtained by deep learning (DL), the notion arises that DL is not agnostic to input errors or biases introduced, in particular in sample-scarce situations. Within the CDL, we seek to understand the influence of different sample quality aspects for creating a sample database: (1) inherited properties (quality parameters of the underlying image such as cloud cover, seasonality, etc.); (2) individual (i.e., per-sample) properties, i.e. data lineage and provenance, (3) reliability of the labelling (classification).

This master thesis explores the influence of samples collected from different camp settings were hand-selected and annotated with computed features in an initial stage. The supervised annotation routine shall be automated in a way that thousands of existing samples can be labelled with this extended feature set. This should help better condition the subsequent DL tasks in a hybrid AI approach.

Suggested reading:

Lang, S., L. Wendt, D. Tiede, Y. Gao, V. Streifeneder, H. Zafar, A. Adebayo, G. Schwendemann and P. Jeremias (2021). "Multi-feature sample database for enhancing deep learning tasks in operational humanitarian applications." GI Forum - Journal for Geographic Information Science 9(1): 209-219. 

Start: 1 Oct 2021

Prerequisites/qualification: remote sensing image classification machine / deep learning

Suggested / supervised by: Stefan Lang, Barbara Hofer, Petra Füreder

Background: The Christian Doppler Laboratory on Earth Observation for Humanitarian Action – GEOHUM – seeks to advance technologies at the interface of EO*GI and AI (artificial intelligence). The applications range from mission planning and operations in crisis intervention to population estimation for food distribution or vaccination campaigns. Our vision is to enhance technical and organisational capacities matching specific needs from humanitarian organisations, in particular our partner MSF (Doctors Without Borders). Envisaged outcomes are a fundamental scientific substantiation and the development and innovative use of relevant information products to optimize aid delivery in conflict and humanitarian disaster situations. Over the last ten years, thousands of dwelling representations in temporary refugee settlements were generated through semi-automated image analysis approaches. This happened by and large demand driven, with a very simplified, often ad-hoc defined underlying data model. Different dwelling types were documented over time by differentiating between traditional or regional dwelling types and artificial dwellings (mainly tents, or other makeshift structures). Within the CDL we try to systematize this inventory by generating an ontology and related data model for dwelling types. It contains (1) a time stamp and unique identifier to the underlying image; (2) geometric properties, (size, orientation, shape), c. spectral features (standardized colour code); (3) context-related properties (arrangement.

This master thesis explores existing schemata and dwelling typologies and supports the development of a purpose-driven domain ontology in the context of humanitarian assistance.  it also includes the investigation of a global grid (e.g., based on Sentinel-2 tiling and subdivisions) for space-time models of dwelling dynamics.

Suggested reading:

Lang, S., L. Wendt, D. Tiede, Y. Gao, V. Streifeneder, H. Zafar, A. Adebayo, G. Schwendemann and P. Jeremias (2021). "Multi-feature sample database for enhancing deep learning tasks in operational humanitarian applications." GI Forum - Journal for Geographic Information Science 9(1): 209-219. 

Start: 1 Oct 2021

Prerequisites/qualification: remote sensingdatabasesontologies