Real Time IP Camera Parking Occupancy Detection using Deep Learning
Albertus Farley, Hanry Ham, Hendra Hendra
Abstract
To find the parking space in a crowded area is quite troublesome due to uncertainty whether the area has an empty parking space or not. If it does not have an empty space, thus it leads to over-stress before doing the main activity and more fuel used. Therefore, We show by only using a IP Camera could allow the parking occupancy detection in real time. The detection were performed by using several deep learning architecture: LeNet, AlexNet, mLeNet, and mAlexNet. The proposed method were validated by training and validating sets of dataset produced combine with the CNRPark + Ext dataset. Results show that the proposed method using mAlexNet with good classification rates up to 93.15% and 0.5 seconds of processing time of each parking space.