Live public parking availability

Data Science , Infrastructure , Non-Profit & Social Impact , Technology

by Craig Nash | Sept. 23, 2018
Utilize cameras with views of public street parking to identify & update the availability in real time using mean-shift algorithm

The idea is to provide a public service which updates a driver in real-time with the location of the closest street parking, or allows them to search parking availability in advance. 

The latest computer learning developments should make this a fairly easy system to design. by feeding color information coming from cameras with views of street parking to a mean-shift algorithmic system that can look for a certain black level or street color, then measure that area to determine if parking is available. Even better options might also exist for determining availability.

 

The biggest challenge would be obtaining the live imaging. This might be possible by searching out public cameras, but might make measurements difficult due to different camera specs. This could be overcome by letting community members "volunteer" cameras by logging the specs in when they grant access, perhaps using something similar to public distributed computing, but with webcams. Another option  could be to physically mount cameras on very tall buildings that can image a large area, since image resolution wouldn't be overly important, just color or reflection levels. Since most in-demand parking is in downtown areas, this could work, though wouldn't be entirely friendly to a free public availability system.