My solution focused on applying Computer vision techniques to the Car Park space. It is said that there are more cameras per inch square overlooking car parks than anywhere else in the city. Yet, none of these car park spaces have cameras overlooking them that can in real time both "count" the number of cars in a car park and provided an updated real time notification of the availability of that space using a mobile app. My main solution was to attempt to use the camera on board a cheap system like the Raspberry Pi camera module to count the number of cars in a space and to let that source be a more reliable source of information to an open source car parking app like ParkJam which has been modified to incorporate that as a source.

Due to time constraints, this project attempted to bring together two open source projects to seamlessly work together using a Raspberry Pi camera back-end for capturing visual data over car parks (using Andrews Sobral open source Background Subtraction Library) to be able to count cars in a space, and notify another open source project like ParkJam of a space availability.

The Ford's OpenXC platform could enhance the experience if it has a means to allow it directly access either simple notification data or video feed covering a car parking spaces. When done in real time, the driver always knows when a desired space is available.

Share this project: