This is an open source code of paper DeepPVMap: Deep Photovoltaic Map for Efficient Segmentation of Solar Panels from Low-Resolution Aerial Imagery.

The project was developed using the following environments.

Install

Install the required dependencies:

pip install -r requirements.txt

Usage

1. Download the dataset and extract it.

   Download Link

2. Set conf.yml model.encoder. Options are:

   • "timm-efficientnet-b0"
   • "timm-efficientnet-b5"
   • "timm-efficientnet-b7"
   • "mit_b0"
   • "mit_b2"
   • "mit_b5"

3. Run the following command to train the model:

   python main.py --name <experiment name> --exp_dir <export directory> --data_dir <data directory>

The map generation script is under development.

Result

Photovoltaic (PV) panels predicted using orthophotos from Taiwan National Land Surveying and Cartography Center (NLSC), collected in June 2022.

Model Checkpoints

Download Link