Annotation of satellite images
Map Communities through Crowdsourcing
Phase-2 : To map the communities in Southeast Bangladesh (Bandarban, Cox’s Bazar, Khagrachhari and Rangamati)
Labelling with place names
A major constraint for infectious disease surveillance and planning healthcare delivery in resource-poor countries is a lack of accurate information on the locations and sizes of settlements in rural areas. In many countries, lists and maps of villages are missing, incomplete and/or out of date. Collecting this information using conventional methods, e.g. in a population census, is resource intensive and expensive is generally done around once every 10 years. We proposed a novel, low-cost citizen science method to support local volunteers to address this problem and Groupmappers has been awarded Public Engagement Bursary funded by Welcome Trust.
The goal of this project is to map settlements in rural Bangladesh through crowdsourcing. The primary objective is to make this data freely available online as it has a wide range of uses beyond health. It will also be used for communicable disease surveillance by the government through our existing collaborative partnership. This project is a follow-on to an unfunded pilot run by our team in Dhaka in late 2017 with a series of workshops. In the pilot, the mapping methods were developed and a 3394 square km area of Southeast Bangladesh was mapped from satellite images by 20 volunteer members using free online resources and the Google Earth platform.
With the Public Engagement Fund we plan to map 4 districts (Bandarban, Cox’s Bazar, Khagrachhari and Rangamati) in the Southeast of Bangladesh. This area is relatively inaccessible and there is an urgent need for accurate maps to guide healthcare provision and numerous other purposes.
The mapping will be done in 5 steps:
- Website, smartphone application and publicity material development – months 1-3
- Annotation of satellite images – months 1-3 (GroupMappers members), months 4-5 (general public)
- Validation of annotation – months 1-5 (GroupMappers members)
- Labelling with place names – months 5-10 (general public)
- Preparation of report and sharing of data – months 11-12 (GroupMappers members)
This seed project will form the basis of a grant application to expand the same platform to map settlements and other features in areas of the Greater Mekong Subregion, India and Bangladesh for use in disease mapping and modelling research and for use by governments in disease surveillance and other applications.
We support communicable disease control programs under the Ministry of Health and Family Welfare (MOHFW) in Bangladesh through collection of high quality, high resolution geospatial data. The initial goal is to map communities in selected rural areas. Once, complete, the next stap will be to label the polygons with the names of communities through crowdsourcing.