Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
531706 | Pattern Recognition | 2006 | 16 Pages |
Abstract
A multi-step recognition process is developed for extracting compound forest cover information from manually produced scanned historical topographic maps of the 19th century. This information is a unique data source for GIS-based land cover change modeling. Based on salient features in the image the steps to be carried out are character recognition, line detection and structural analysis of forest symbols. Semantic expansion implying the meanings of objects is applied for final forest cover extraction. The procedure resulted in high accuracies of 94% indicating a potential for automatic and robust extraction of forest cover from larger areas.
Keywords
Related Topics
Physical Sciences and Engineering
Computer Science
Computer Vision and Pattern Recognition
Authors
Stefan Leyk, Ruedi Boesch, Robert Weibel,