Scaling Machine-Learning Workflows for Participatory Mapping: Monrovia ML Challenge
Building on the work conducted during the Open Cities Africa Monrovia project (2018-2019), HOT and iLab Liberia will contribute to the development of operational machine learning (ML) methods to improve both the scale and speed of the participatory mapping workflow.
As part of this project, HOT and iLab Liberia conducted the following activities contributing towards machine learning (ML) development:
- Capturing UAV imagery for digitization and ML algorithm development
- Capturing and using StreetView imagery for ML feature identification/extraction
- Exploring AI-assisted mapping in OpenStreetMap with roads and buildings.
Through these activities, as well as other ML projects that HOT is involved with, HOT and iLab Liberia have documented lessons learned and recommendations for the integration of ML into the participatory mapping workflow.
This project will be part of a larger international ML challenge, inviting the best ML practitioners to develop open source algorithms and workflows which can be used to generate further geospatial data for the city using automated techniques. The project provides the opportunity for local teams to learn about machine learning concepts and applications; to receive training on applying these methodologies locally; and to collaborate with an extended team investigating how this technology can potentially speed up the mapping process, through combining human skills with computer automation.
Recent news from Scaling Machine-Learning Workflows for Participatory Mapping: Monrovia ML Challenge (View all news)
HOT and iLab Liberia Exploring the Potential of Machine Learning to Augment Human Mappers in Monrovia
HOT worked with iLab Liberia in Monrovia to test the potential of machine learning algorithms to automate repetitive mapping tasks and highlight areas in need of updating to maximize the impact of human mappers.