Smartphone could help make better decisions on solving road infrastructure problems.
Maintenance departments need to
regularly assess the quality of the roads in order to
properly maintain them. Currently, this is done by yearly
inspections or in response to reports from the general
public. It would be advantageous to continuously monitor the
road surface so that damages like rutting and potholes can
be detected as soon as they occur. Furthermore, detection of
precursor signs like cracks will allow the maintenance crews
to address problem areas before they develop into serious
problems. We want the system to be inexpensive and easy to
run.
Our approach is to collect images, GPS and other data with
smartphones, use computer vision algorithms to analyze the
images, and save the results in the database of the
maintenance department where it can be displayed to the user
or further analyzed. Additional sensors and devices like
OBDII or structured light sensors could be added to get
additional data.
Our pilot project is partnered with City of Pittsburgh. The City uses our system, integrates it in their workflow
and evaluates its effectiveness.
The same dataset can also be used to find and asses traffic signs. We have developed and implemented a method that detects and evaluates stop signs. Signs in poor condition were automatically detected: vandalized by sticker or graffiti, occluded by vegetation, or displaced.
We are part of the Robotics Institute, School of Computer Science, Carnegie Mellon University.
We are associated with Traffic21 and UTC.
Acknowledgement >>