More information: http://www.directionsmag.com/article.php?article_id=2457&trv=1
RoadTracker is a new tool written by GeoEye as a plugin to Feature Analyst. RoadTracker provides tools that allow for quick and easy semi automated extraction of features such as roads, streams and trails. The user needs to digitize the feature by clicking only on the end points - RoadTracker will then automatically extract the feature by analyzing and tracking the feature from the start to end point. On roads and trails that have a lot of curves in them - in my opinion - RoadTracker will save the analyst upwards of 60% of extraction time. In addition, it will also reduce errors that get introduced due to user fatigue while digitizing complex features.
RoadTracker Key Features
- High degree of automation speeds centerline detection
- Automatic detection of intersections and adjustment of existing centerlines improves network accuracy
- Automatic centerline feature attribution based on imagery analysis removes manual rework
- 'Smart editing' simplifies feature correction workflow
- Complements any geospatial environment that supports VLS' Feature Analyst
RoadTracker is a proprietary application built on top of Feature Analyst. It is a semi-automated tool that detects road centerlines in orthophotos and automatically computes road attribute data when the centerlines are extracted as vector features. The tool changes the traditional workflow for centerline creation from manually intensive point-and-click digitizing to an efficient process that automates centerline creation, attribution, intersection insertion and topology cleaning. - directionsmag
(from DirectionsMag)
After selecting a RoadTracker editing tool from within a standard geospatial package, the user collects two pixels that approximate a new centerline's endpoints (A). RoadTracker creates an initial solution (B). Existing road features are suppressed so that when the same two endpoints are used, a new path is found that completes the loop (C). (Click for larger image)
No comments:
Post a Comment