• GPS Lands Singapore

    Mobile Mapping to Create Autonomous Vehicle Maps in Singapore

    Singapore

Project Summary

Project

Singapore’s wide-ranging Smart Nation project seeks to incorporate innovative technology into various aspects of daily life. As part of the project, Singapore’s Land Transport Authority and Jurong Town Corporation designated the One-North tech hub as a trial area for testing autonomous vehicles (AV). However, to ensure safety, the project needs accurate information about all aspects of the test environment. GPS Lands Singapore was contracted to produce highly detailed maps of routes totaling 55 kilometers to help AVs navigate.

Solution

Before mapping began, GPS Lands Singapore used satellite images and traditional photographs of streets to get a first impression of the site and plan routes for the mapping process. The organization then deployed a car-mounted laser scanning system to capture image data of the roads and surrounding elements of the test environment.

Outcome

The process of planning, image capture, and data extraction took four months. By adjusting auto-detection tools within Orbit Feature Extraction and filtering out false detections, GPS Lands Singapore reduced the time needed to manually search for specific features within the data. The completed AV guidance map will be used to help other map providers or AV companies develop their own AV mapping data to suit specific needs.

Software

GPS Lands Singapore used Orbit applications to capture and process all acquired image data. After the team generated point cloud data, Orbit Feature Extraction automatically identified road features such as lane markings and curb lines. Detecting traffic signs and poles helps operators determine AV driving behavior and accepted destinations along the test routes.

Project PlaybookOrbit Content Manager, Orbit Feature Extraction

Outcome/Facts
  • As part of Singapore’s Smart Nation project, the country designated 55 kilometers of roads for testing autonomous vehicles.
  • To help autonomous vehicles navigate safely, the project needs highly-detailed maps of driving routes.
  • After the team generated point cloud data, Orbit Feature Extraction automatically identified road features such as lane marking and curb lines.
  • The completed AV guidance map will be used to help other map providers or AV companies develop their own AV mapping data to suit specific needs.