Spatial resolution requirements for automated cartographic road extraction
Ground resolution requirements for detection and extraction of road locations in a digitized large-scale photographic database were investigated. A color infrared photograph of Sunnyvale, California was scanned, registered to a map grid, and spatially degraded to 1- to 5-metre resolution pixels. Road locations in each data set were extracted using a combination of image processing and CAD programs. These locations were compared to a photointerpretation of road locations to determine a preferred pixel size for the extraction method. Based on road pixel omission error computations, a 3-metre pixel resolution appears to be the best choice for this extraction method. -Authors
Citation Information
Publication Year | 1990 |
---|---|
Title | Spatial resolution requirements for automated cartographic road extraction |
Authors | S. Benjamin, L. Gaydos |
Publication Type | Article |
Publication Subtype | Journal Article |
Series Title | Photogrammetric Engineering and Remote Sensing |
Index ID | 70016191 |
Record Source | USGS Publications Warehouse |