Image interpretation — inferring land use to map farmland and woodland in Bangladesh for community resilience
What is image interpretation?
Image interpretation is a term describing various techniques that are used to infer characteristics about real-world objects or places from images taken of them.
A correctly interpreted image can be used to infer lots of information about those real-world objects — it’s possible to measure how big or small something is, what it might be made of, or how often its used, for example.
Image interpretation is the foundation of mapping for Missing Maps: a mapper looks at a bird’s eye view of an area (most likely an image taken from a satellite), they then interpret what they see in the image and then trace digital information on top of it which is added to OpenStreetMap (find out more about this here).
Although this blog focuses on some training we provided to mappers in Bangladesh, it gives an insight into the kind of image interpretation techniques that are used all over the world, to map all sorts of things, in all sorts of contexts.
Context
In 2018, we were asked to provide remote information management support to a Vulnerability and Capacity Assessment (VCA) that was happening as part of the Vulnerability to Resilience (V2R) programme in Teknaf, Bangladesh.
The programme was a partnership between the Bangladesh Red Crescent Society (BDRCS) and British Red Cross to support community resilience through a multi-sector approach.
At the start of the programme, communities in Teknaf were being heavily impacted by an increase in Rohingya population who had been fleeing instability in neighbouring Myanmar. The influx was leading to increased tensions between host communities and Rohingya, due to competition for resources to cover basic daily needs. For example, competition for firewood and a diminishing water supply was degrading the environment, as well as competition for jobs and a reduced access to healthcare facilities.
The resentment from host communities also increased due to their perception that the Rohingya population was receiving humanitarian assistance whilst host communities were neglected.
The V2R Programme aimed to mitigate the impact of the issues in Teknaf through strengthening resilience of affected host communities to reduce tensions between those communities and the Rohingya population.
The programme focussed on mainstreaming protection, including enhancing awareness of human trafficking, and supporting BDRCS to better respond to emerging risks and hazards.
What did we do?
Part of the VCA involved mapping of farmland and woodland resources in the project area. The majority of the mapping was undertaken by Bangladesh RC volunteers — they already had experience mapping from satellite images, but only of buildings and roads, so wanted some guidance about what to look for to map the farms and woods.
We had access to some photographs of farmland and woodland that were taken by someone working on the programme in the field. We used these in conjunction with some of the satellite images they would be using to prepare a guide that showed what features on the ground looked like from the air.
What was the result?
The guide gave the remote mappers the understanding and confidence to interpret the images correctly and add the data needed to the map. In total, 72 different contributors made 6,286 edits to OpenStreetMap and added the full extent of farmland and woodland in the area. This enabled these two community resources to be properly assessed and valued as part of the VCA process.
Impacts of the V2R Programme so far
Although the bulk of our team’s support to V2R was in 2018, the programme itself is still running and is going from strength to strength. It’s beyond the scope of this blog to detail all the successes, but in brief:
1200 households have received conditional cash grants to reinforce their livelihoods, based on business plans developed under V2R
1212 beneficiaries have received technical training on livelihoods and financial literacy, with a focus on maintaining cash books and understanding of profit margins
Sea safety kits consisting of eleven items such as a radio, torchlight, signal torch, compass, life jacket, and first aid box were distributed to 599 fishermen to ensure their safety at sea
The programme has also shown its adaptability to the new challenges brought about by Covid-19:
The project trained 29 community tailors and produced 29790 face masks and 3162 households received face masks which were produced by those community tailors
And from an Information Management perspective:
129 project staff and unit volunteers received training on mobile data collection, with an additional 115 staff and volunteers trained on mobile data collection to conduct assessments and household surveys
Want to know more about image interpretation?
There’s a lot of diversity in the use of image interpretation, so if you’d like to know more about how it’s used in conjunction with other techniques to infer information, I think these articles are worth a read:
What Oil, Satellite Technology and Iraq can Tell us About Pollution — Bellingcat
Talc: the everyday mineral funding Afghan insurgents — Global Witness
