Abstract
In this research work we proposed secure parking framework based on video analytics for human activity recognition and user interaction through smartphone application. We used machine learning algorithms for human activity recognition using smart camera. Algorithms were used to process the parking lot video to extract meanings form human activities while mobile phone application was used to communicate with the user. Image registration algorithm marked incoming and leaving vehicles to keep track of available parking bins. Activities happening in the parking area e.g. human-vehicle interaction, human-human interaction, parking, entering and exiting the parking lot were recognized using support vector machine classification on space time features representation of the scene. In order to train and test the activity identification algorithm we used real world video dataset VIRAT (1.0). Video analytics algorithms were embedded into smart camera hardware. Video backup was maintained at backend surveillance server.