Building Footprint Extraction - China

This deep learning model is used to extract building footprints from high-resolution (15–25 cm) imagery. Building footprint layers are useful in preparing basemaps and analysis workflows for urban planning and development. They are also used in insurance, taxation, change detection, infrastructure planning, and a variety of other applications.

Digitizing building footprints from imagery is a time-consuming task and is commonly done by digitizing features manually. Deep learning models have a high capacity to learn these complex workflow semantics and can produce superior results. Use this deep learning model to automate this process and reduce the time and effort required for acquiring building footprints.

Using the model

Fine-tuning the model
This model can be fine-tuned using the Train Deep Learning Model tool. Follow the guide to fine-tune this model.

Input
8-bit, 3-band high-resolution (15–25 cm) imagery. 

Output

Feature class containing building footprints.

Applicable geographies

The model is expected to work in urban areas of China.

Model architecture

The model uses the MaskRCNN model architecture implemented using ArcGIS API for Python.

Accuracy metrics

The model has an average precision score of 58 percent. 

Training data
This model was trained on BONAI dataset.

Sample results
Here are a few results from the model.

Citations

Wang, J., Meng, L., Li, W., Yang, W., Yu, L., & Xia, G.-S. (2022). Learning to extract building footprints from off-nadir aerial images. IEEE Transactions on Pattern Analysis and Machine Intelligence, Advance online publication. https://doi.org/10.1109/TPAMI.2022.3162583