CurvePortal for Automated Identification and Extraction of Horizontal Curve Information

This project will develop a prototype of CurvePortal, a web interface for extracting horizontal curve location and geometric information automatically from geographic information system (GIS) roadway maps with improved curve data extraction algorithm. Work in Stage 1 will focus on incorporating and validating MIRE (Model Inventory of Roadway Elements) compatibility and improving the curve data extraction algorithm, especially increasing its accuracy when applied on low-quality GIS roadway maps. The extracted curve data will include MIRE curve types as well as all MIRE horizontal curve elements, except for superelevation which can only be obtained from field survey. To improve the curve data extraction algorithm, addition of artificial intelligence components to the algorithm will be explored to address the centerline deviation issue. Other methods, such as image processing-based methods, will be explored to identify spaced-out centerline segments and improve the resolution of the GIS centerline feature. In addition, calibration automation, given the availability of ground truth data, will be added to the application and tested. Manual work will be performed to obtain sets of ground truth data for calibration (possible source can be aerial/satellite images). Work in this stage will focus on developing a prototype of CurvePortal, a web application which can be accessed by transportation agencies to upload their GIS roadway shapefiles for curve data extraction. CurvePortal will be developed with user interface for transportation agencies to upload their GIS files, and enter necessary information about the data. The required data is expected to include road classes targeted for curve data extraction, the road direction field, and ground truth data. The web application will have a server with a database to store the data uploaded and information entered. The curve extraction will be performed by the CurvePortal server program which will be developed to complete the following tasks: (i) automatically preprocess the uploaded GIS roadway maps; (ii) automatically detect and group uploaded maps with similar vertex patterns (e.g., average distance between vertices at curves) to prepare for calibrating the CurveFinder algorithm. If ground truth data is not provided, manual work will be performed to select sample aerial/satellite images that contain 50 curves and obtain the ground truth curve information for each group of GIS maps with similar vertex pattern. Automated methods, such as those based on image processing, will be explored to replace the manual methods; (iii) automatically calibrate the CurveFinder algorithm based on the ground truth; and (iv) automatically perform curve identification and curve information extraction and create a curve GIS shapefile and make the shapefile downloadable by the client. Toward the end of this stage, technology transfer efforts will be initiated in collaboration with state departments of transportation to test and validate all functions of CurvePortal including both the web application functions and the server program functions. The CurvePortal prototype system can be further evaluated by the Federal Highway Administration (FHWA). If successful, plans for nationwide implementation will be developed.

Language

  • English

Project

  • Status: Active
  • Contract Numbers:

    Project 20-30, IDEA

  • Sponsor Organizations:

    Federal Highway Administration

    1200 New Jersey Avenue, SE
    Washington, DC  United States  20590

    American Association of State Highway & Transportation Officials (AASHTO)

    444 North Capitol Street, NW, Suite 225
    Washington, DC  United States  20001

    National Cooperative Highway Research Program

    Transportation Research Board
    500 Fifth Street, NW
    Washington, DC  United States  20001
  • Project Managers:

    Jawed, Inam

  • Performing Organizations:

    University of Wisconsin, Madison

    Madison, WI  United States  53706
  • Principal Investigators:

    Noyce, David

    Chitturi, Madhav

  • Start Date: 20150715
  • Expected Completion Date: 0
  • Actual Completion Date: 0
  • Source Data: RiP Project 40034

Subject/Index Terms

Filing Info

  • Accession Number: 01569558
  • Record Type: Research project
  • Source Agency: Transportation Research Board
  • Contract Numbers: Project 20-30, IDEA
  • Files: TRB, RiP
  • Created Date: Jul 16 2015 1:00AM