SQUID – automating the detection of #potholes in #NYC
Posted on March 28th, 2016
03/28/2016 @ UrbanFutureLab, 15 MetroTech, 19th fl, Brooklyn, NY
Varun Adibhatla @Advanced Research in Government Operations (@ARGO) talked about their initiative to more quickly map the potholes in NYC’s 6000 miles of roads.
Monitoring is currently done by a dedicated group (Scout – street condition observation unit) that does “windshield surveillance” of all roads. Varun’s group is working an automated system which sends accelerometer data and pictures to the cloud. The accelerometer indicates street hot spots and the images can be scanned or serve as verification of the problem.
SQUID (street quality identification device) is run by a Raspberry Pi that is attached inside the trunk with a camera mounted next to the backup camera. Data are streamed to the cloud with one 15kb per image * one image per second. In their trials they drove 400+ miles in a week and collected 10 gig of data.
Next Varun showed an interactive map powered by Tableau that highlights the bumpy routes and allows one select images taken of the road.
Finally he talked about future initiatives including
- Anticipatory maintenance
- Using microphones to records street noise as a classifier for potholes
- A smartphone app that crowd sources street condition data to the cloud
- Adding air quality sensors to provide a more complete picture of the environment
He recommended the book ‘Smart Cities’ by Anthony Thompson as a consumer handbook for smart cities globally.