Help users find available parking spots in a specific area. Implementation of a backend software interface for a parking detection system that uses MASK R-CNN neural network and additional logic based on the segmentation output of the network.
The search for parking in big cities over the world is becoming a more and more difficult mission. A US Today research(reference in project book) shows that an average New Yorker spends 107 hours a year in search for parking, and the average US citizen 17 hours per year. Another research made by Calcalist newspaper (reference in project book) shows that in Israel big cities the parking search time has grown by 17% to 80% between 2012 to 2016 (the growth varies between cities and time of day). This project implements a system that will turn the uninformed search for parking to an informed search. This change will help reduce the search time by giving the drivers the parking spots status near their destination in real-time. By using Mask R-CNN output statistics, the system will map the parking lots in every undercover street. This map will be done offline. After the mapping for those streets, the system will get images streams from different platforms and update the s tatus of those parking lots. Every driver will be able to get the status of his surrounding undercover streets and know where he has the best chance to find parking.
