Searching for shelters
Locating internally displaced people... one tent at a time
In Afghanistan, there are almost 2.6 million IDPs. In 2018, a severe drought in the north of the country led to an additional 371,000. Many are living in informal tented communities. These tents are important indicators in determining the changing demand for humanitarian support.
The Norwegian Refugee Council (NRC) came to Alcis asking for help in locating the missing people, who had moved into temporary camps due to drought forcing them from their homes.
Initially we looked at the impact of the drought, identifying the communities where the loss of crops was most felt. More recently we have been looking at the resultant movement of people; analysing the location and distribution of shelters in makeshift camps.
We used a combination of Satellite imagery and machine learning to detect and count temporary IDP shelters in Afghanistan. Machine learning and Geographic Information Systems were used to identify individual IDP tents in Afghanistan for humanitarian shelter.
After recognising that the tents provided by NRC have a unique signature when observed in satellite imagery, we were able to see the rate at which IDP camps appeared and evolved in a very short span of time.
Our results displayed 99% accuracy in identifying tents.
How we did it
Our team of experts are here to help develop and deliver the technical components of your requirements, including:
Programme design & Implementation
Monitoring and Evaluation
Organisational development (capacity building, mentoring and training)
National Spatial Data Infrastructure support, assessment and delivery
Academic research and methodological innovation