USAFORUNHCRhive/turkana-camp-roof-mapping

# Introduction **Jump To:** Setup | Data | Examples | Acknowledgments | License | Contributing Building and solar panel identification in aerial imagery for humanitarian mapping. ***Fig 1.** Drone imagery over Kalobeyei Settlement. Some solar panels and buildings with metal sheet , plastic , thatch , and other/unknown roofing are annotated.* This project automates the identification of buildings and solar panels in aerial imagery, aiding humanitarian mapping. In many developing regions, maps are often outdated or missing, hindering development planning and disaster response. Our work accelerates the creation of up-to-date maps, offering valuable insights into the infrastructure of these areas. This repository houses models designed for two primary tasks: 1. Detecting buildings and solar panels in aerial imagery. This information can be crucial for infrastructure development planning and humanitarian interventions, as well as for monitoring the adoption of renewable energy sources like solar panels. 2. Classifying the roof materials of buildings. This is particularly important as the type of roofing material can provide information about the building's structure, its age, and even socio-economic indicators. # Setup To setup the project, follow these steps: 1. Clone the repository and navigate into the project directory. 2. Create and activate a new conda environment using the provided environment.yml file. ```bash cd conda env create -f environment.yml conda activate ``` 3. Install this repo as a package for easier code imports. ```bash cd pip install -e . ``` 4. (Optional) Use configuration files and a `.env` file for easier script execution. The scripts in this repository take command-line arguments, but it's often easier to store these arguments in configuration files. Example configuration files are provided in `config/`. For path-related arguments, the example configuration files contain paths relative to this …