Trevor Linton’s research while affiliated with University of Utah and other places

Raster map images (e.g., USGS) provide much information in digital form; however, the color assignments and pixel labels leave many serious ambiguities. A color histogram classification scheme is described, followed by the application of a tensor voting method to classify linear features in the map as well as intersections in linear feature networks. The major result is an excellent segmentation of roads, and road intersections are detected with about 93% recall and 66 % precision.

The automatic classication of semantic classes (background, vegetation, roads, water, political bound- aries, iso-contours) in raster map images still poses signicant challenges. We describe and compare the results of three unsupervised classication algorithms: (1) k-means, (2) graph theoretic (GT), and (3) ex- pectation maximization (EM). These are applied to USGS raster map images, and performance is measured in terms of the recall and precision as well as the cluster quality on a set of map images for which the ground truth is available. Across the six classes studied here, k-means achieves good clusters and an average of 78% recall and 70% precision; GT clustering achieves good clusters and 83% recall with 74% precision, Finally, EM forms very good clusters and has an average 86% recall and 71% precision.

The extraction of semantic features from images of geographic maps is a difficult and interesting problem. Such features may be robustly segmented through the use of Gestalt principles such as similarity and continuity as realized through the use of tensor voting methods and color histogram analysis, respectively. A framework is developed implementing these Gestalt principles through various algorithms. Linear feature segmentation and intersection detection methods are given, and their performance is demonstrated on a set of real and synthetic map images. Master of Science;