Geoinformation for Humanitarian Aid
What we do
A pivotal supplementation for official geodata and remote sensing data, user-generated spatial data such as those provided by the OpenStreetMap project or the Social Web are becoming ever more important for efficient crisis management. By mashing different datasets, the resulting situation intelligence enables humanitarian relief organizations and first responders to quickly gain awareness of the situation in the aftermath of a disaster event.
This information also plays a pivotal role in the efforts to mitigate risks, prepare for events, and limit their potential effects. In cooperation with users and relief workers, we are developing innovative processes and services for new analysis methods to better use this potential of different data sources.
Information about the distribution of the population in an area is of pivotal importance for humanitarian aid. The distribution of buildings can serve as a proxy. Using the smartphone app MapSwipe, even users without prior experience can identify populated areas on a satellite image. MapSwipe was developed as part of MSF’s (Médicins sans Frontières/Ärzte ohne Grenzen e.V.) Missing Maps Project, and provides a valuable groundwork for the subsequent geodata gathering in the areas. For analyzing and visualizing the user-generated data, we provide the web service MapSwipe Analytics.
Our application MapSwipe Live shows the newest submissions from MapSwipe App users. That way, project managers can swiftly respond to problems during the data collection. To improve the user experience, we display each user’s own submission in an easy-to-understand rendering.
How do I do it right?
Detecting houses and other infrastructure on a satellite image sounds like a simple task – an impression reality does not always support. We have developed MapSwipe Tutorial, a service providing assistance in unclear situations.
Use MapSwipe Data for Your Projects
The MapSwip dataset contains several million classifications for different geographic regions. We provide services to enable relief organizations and other interested parties to use this vast pool of information: MapSwipe Processing offers download features. Alternatively, the data can be directly included in your application through our geoserver.
The routing service enables users to plan routes based on up-to-date OpenStreetMap data, all while taking into account disaster-related road conditions. This way, the on-site response teams always have access to continuously updated information about reachability and navigability of the surrounding roads.
Avoid Blocked Roads
After the 2015 earthquakes in Nepal, the global OpenStreetMap community mapped up to 800 road kilometers – every hour! This included information about blocked and impassable roads. We render these data usable for evacuation and response planning.
We currently provide hourly update intervals for geodata of all of Africa, South America, and South Asia. During larger disasters affecting regions not yet covered by the service, the addition of new areas can be requested.
Nowadays, Machine Learning and Deep Learning approaches are steadily gaining popularity within the humanitarian (mapping) community. With our cooperating partner team at GIScience Research Group Heidelberg University we investigated the potential of Deep Learning in combination with MapSwipe’s crowdsourcing approach. To this end, we propose a novel workflow to combine deep learning (DeepVGI) and crowdsourcing (MapSwipe). Our strategy for allocating classification tasks to deep learning or crowdsourcing is based on confidence of the derived binary classification.
Deeply learning from satellite imageries
Deep learning techniques, esp. Convolutional Neural Networks (CNNs), are now widely studied for predictive analytics with remote sensing images, which can be further applied in different domains for ground object detection, population mapping, etc. These methods usually train predicting models with the supervision of a large set of training examples.
Fill data gaps
The scarce availability of accurate and up-to-date human settlement data remains a major challenge, e.g., for humanitarian organizations. We investigate the complementary value of crowdsourcing and deep learning to fill the data gaps of existing earth observation-based (EO) products.
Interaction between human beings and machines
VGI data from OpenStreetMap and the mobile crowdsourcing application MapSwipe which allows volunteers to label images with buildings or roads for humanitarian aids are utilized. We develop an active learning framework with deep neural networks by incorporating both VGI data with more complete supervision knowledge.
The Critical Numbers Tool visualizes the mapping progress for the tasks in the Tasking Manager of the Humanitarian OpenStreetMap Team (HOT).
A service providing real-time OSM data.