What question did this study set out to answer?

Develop an automated framework combining deep learning segmentation and polygon regularization to extract structured building footprints from aerial imagery and reduce manual cartographic labor.

December 17, 2019IEEE Transactions on Geoscience and Remote Sensing

Toward Automatic Building Footprint Delineation From Aerial Images Using CNN and Regularization

View Full Paper

Ask AI

Bookmark

Develop an automated framework combining deep learning segmentation and polygon regularization to extract structured building footprints from aerial imagery and reduce manual cartographic labor.
Engineered a scale-robust fully convolutional network (FCN) utilizing multi-scale feature pyramid aggregation combined with post-processing refinement strategies.
Implemented a two-stage coarse and fine polygon regularization algorithm to convert raster segmentation outputs into clean geometric footprints.
Evaluated performance on diverse aerial datasets, including an open dataset comprising 181,000 buildings across varying styles and resolutions.
Automated delineation successfully replaced manual cartographic tracing for at least 50% of individual buildings within the test area.
The multi-scale FCN segmentation outperformed several recent benchmark segmentation models.
The polygon regularization algorithm maintained robust boundary refinement under challenging conditions, including varying image resolutions and low-quality segmentations.

Discussion

No takes yet. Share an insight, caveat, or question.

Implication

Computational study demonstrates automated building footprint extraction across aerial imagery, indicating potential to cut manual mapping efforts by half.

Cite This Study

Wei et al. (2019) studied this question.

synapsesocial.com/papers/69e2b003255fe7814cb73718 https://doi.org/10.1109/tgrs.2019.2954461

View Full Paper

Ask AI

Bookmark

Implication

Computational study demonstrates automated building footprint extraction across aerial imagery, indicating potential to cut manual mapping efforts by half.

Toward Automatic Building Footprint Delineation From Aerial Images Using CNN and Regularization

Key Points

Discussion

Implication

Cite This Study

Toward Automatic Building Footprint Delineation From Aerial Images Using CNN and Regularization

Key Points

Discussion

Implication

Cite This Study

Also Consider

Also Consider