This is a diary about a recent drone mapping initiative that I participated in. I would like to give credits to the Open Mapping Hub - West and North Africa, OpenStreetMap Sierra Leone, Pete Masters and Ivan Gayton for all the support and knowledge given during the course of this initiative. I hope to learn more as we continue collaborating.
Purpose
The use of satellite imagery from multiple sources has been a pivotal aspect of open mapping campaigns across the world. However, satellite imageries have some limitations, such as low resolutions and delayed visitation time. This affects the quality of the digitization of physical features that are to be mapped. In order to address these limitations in open mapping campaigns, the use of Unmanned Aircraft Systems such as drones have been employed to capture images with high resolutions within desired timeframe.
Scope
With support from the West and North Africa Hub through the Mwalai microgrant, OpenStreetMap Sierra Leone embarked on the collection and processing of drone imageries in three slums across Freetown as part of the Know Your City initiative. The imageries would be used to map buildings and critical infrastructure, test the fAIr model and Field Mapping Tasking Manager.
Technical Specifications/Parameters
In order to have high resolution imagery, the team used a DJI Mavic 2 Zoom drone for the flights and Open Drone Map for the image processing. A smartphone-based flight planning and control application was used to conduct flights with specific elevation, overlap, and angle settings, allowing the operators to ensure consistent resolution, quality, and coverage across the areas of interest.
Flights were conducted above slums with dense buildings, therefore, flight settings with paths and gimbal angles that captured the top and sides of buildings and other infrastructure. The team used the following flight settings. 70% Frontal Overlap 80% Side Overlap -75 degrees Gimbal Angle
Processing
After collecting the images, we use WebODM, Open Drone Map web version to stitch the images and create orthophotos and other data products. We use the following processing settings and outputs in the WebODM interface: Auto-boundary: yes dsm: yes dtm: yes pc-quality: high
Tasking Manager
The processed imageries were uploaded to the HOT Tasking manager as basemap for various tasks to map buildings. The Tile Mill Server (TMS) links to the imagery sets were added to the task instructions for mappers to use to help them distinguish the boundaries of building footprints.
Buildings Mapped
At the end of the campaign, (number of buildings) were mapped across 3 slums and the buildings were used as trials for the Field Mapping Campaign through the Field Mapping Tasking Manager to conduct household surveys.