Transfer Learning with Image Data Augmentation for Parking Occupancy Detection

The issue of parking in the city is more complex than it appears. In the urban context where space is limited, an efficient parking management is more required. Solutions are counter-intuitive and require integrated parking management to detect space of parking. Deep learning is one of the branches of artificial intelligence, which has been widely used as an effective and robust technique to classify images associated with parking images. Transfer learning is a technique based on deep learning that can be used for image classification. In this paper, we exploit the Inception V3 model in the context of transfer learning technique with image data augmentation to classify the parking space of PKLot dataset. The experimental part shows interesting results in term of accuracy and the number of computational parameters.