The problem SafePark solves

Illegal parking in cities eats road space away, leading to frequent traffic jams. Vehicles are often parked haphazardly without any consideration for other commuters. In just a short span of time, it reduces the traffic bearing capacity of the road, hampers public transport systems, and contributes to traffic jams.

Solving the issue of traffic congestion has multiple direct and indirect impacts - travel time reduction, saving commuters’ fuel costs, air and noise pollution reduction, preventing potential accidents, etc. The problem of illegal parking in these high traffic density roads has a domino effect over time as volumes of vehicles pile on behind this gridlock - a simple problem which can easily be resolved. Reducing road encroachment is a step in the right direction to improve the liveability index of metropolitan cities, such as Bengaluru.

Challenges we ran into

1. Agrregating a large, reliable and good quality dataset to train models that can precisely track/segment/mask cars and road lanes in real time as well as all kinds of lighting.
2. Hardware requirements to train/run these models.
3. Coming up with an algorithm to detect amount of overlap and calculate the true operational area of the road beacuse of each encroaching vehicle.