Skip to content

Detection Class: Roads

In Development


Model BSHRK-W-RDD-09 is trained to detect roads. BSHRK-W-RDD-09 is capable of detecting roads from aerial image input and output a high-quality map of road coverage as a binary pixel mask.

Expected Output

When the above input specifications are met, BSHRK-W-RDD-09 will produce an image mask as a GeoTIFF file with a resolution matching that of the input imagery.

The results are provided as a binary pixel mask indicating road coverage.

For output troubleshooting and limitations, see Known Issues and Limitations below.

Road Detection

For accurate road detection the inference process is first completed. Each pixel is given a value for the likelihood of being the target object (road). After the inference process, the road segmentation process is completed.

Road segmentation uses the grayscale data output by the inference process and evaluates it against a set threshold value. Any pixel with a value higher than the threshold value is assigned a value of 1 (road); any pixel below the threshold value is assigned a value of 0 (not road).


Let us know that you're looking for and we would love to hear from you!

Blackshark Inquiries