Automatic and Accurate Extraction of Road Intersections from Raster Maps

Abstract

Since maps are widely available for many areas around the globe, they provide a valuable resource to help understand other geospatial sources such as to identify roads or to annotate buildings in imagery. To utilize the maps for understanding other geospatial sources, one of the most valuable types of information we need from the map is the road network, because the roads are common features used across different geospatial data sets. Specifically, the set of road intersections of the map provides key information about the road network, which includes the location of the road junctions, the number of roads that meet at the intersections (i.e., connectivity), and the orientations of these roads. The set of road intersections helps to identify roads on imagery by serving as initial seed templates to locate road pixels. Moreover, a conflation system can use the road intersections as reference features (i.e., control point set) to align the map with other geospatial sources, such as This work is based on an earlier paperroad orientations. For the problem of road intersection extraction, our approach achieves over 95% precision (correctness) with over 75% recall (completeness) on average on a set of 70 raster maps from a variety of sources.

Efficient and Robust Graphics Recognition from Historical Maps

“…Figure 2 illustrates the different steps and the user interface of the described technique. Our graphic recognition approach consists of three major steps: (i) separation of homogeneous thematic map layers using color image segmentation (CIS) [Leyk, 2010], (ii) interactive extraction and cleaning of these separated map layers, and (iii) subsequent raster-to-vector conversion of the cleaned map layers [Chiang et al, 2008;Chiang and Knoblock, 2011].…”

“…Once the road layer has been cleaned, we employ our previously described technique for automatic generation of road geometry [Chiang et al, 2008] and subsequently convert the road geometry to road vector data automatically [Chiang and Knoblock, 2011]. Figure 2(j) shows the road vectorization results (without manual post-processing) consisting of 1-pixel wide road centerlines.…”

“…We demonstrate this approach using a historical U.S. Geological Survey (USGS) topographic map in a road vectorization example, which builds on our previous work on color image segmentation (CIS) [Leyk, 2010] and road vectorization [Chiang et al, 2008;Chiang and Knoblock, 2011]. The USGS topographic map suffers from poor graphical quality and thus represents a particularly challenging research object.…”

A Method for Automatically Extracting Road Layers from Raster Maps

“…For example, the scanned USGS topographic maps can be downloaded from the Microsoft Terraserver and many other information rich raster maps can be found in map repositories such as the University of Texas Map Library. 1 To utilize the information in the raster maps, in our previous work, we developed a technology to first identify the road intersection templates in the raster maps [4] and then match the set of road intersection templates with another set of road intersection templates from a georeferenced data set (e.g., vector data) [2] to identify the geocoordinates of the maps and align the maps with the georeferenced data. For the automatic road intersection extraction process, in [4], we employed a histogram analysis approach to extract the foreground pixels from the raster maps and utilized a text/graphics separation 1 http://www.lib.utexas.edu/maps/ algorithm [1] to extract the road lines.…”

“…1 To utilize the information in the raster maps, in our previous work, we developed a technology to first identify the road intersection templates in the raster maps [4] and then match the set of road intersection templates with another set of road intersection templates from a georeferenced data set (e.g., vector data) [2] to identify the geocoordinates of the maps and align the maps with the georeferenced data. For the automatic road intersection extraction process, in [4], we employed a histogram analysis approach to extract the foreground pixels from the raster maps and utilized a text/graphics separation 1 http://www.lib.utexas.edu/maps/ algorithm [1] to extract the road lines. However, due to the complexity of maps and noise introduced in producing the maps in raster format (e.g., scanning), it is difficult to separate the foreground pixels from raster maps automatically; even manual work requires tedious work to select the colors that represent the foreground pixels in a scanned map.…”