Deep Learning for Decentralized Parking Lot Occupancy Detection

This repo contains code to reproduce the experiments presented in Deep Learning for Decentralized Parking Lot Occupancy Detection.

Visit the project website for more info and resources (dataset, pre-trained models).

Requirements

• Caffe with Python interface (PyCaffe)

Steps to reproduce experiments

1. Clone this repo together with its submodules:

   git clone --recursive https://github.com/fabiocarrara/deep-parking.git
   

2. Download the datasets using the following links and extract them somewhere.

   Dataset	Link	Size
   CNRPark	http://cnrpark.it/dataset/CNRPark-Patches-150x150.zip	36.6 MB
   CNR-EXT	http://cnrpark.it/dataset/CNR-EXT-Patches-150x150.zip	449.5 MB
   PKLot	visit PKLot webpage	4.6 GB

3. Get the dataset splits and extract them in the repo folder

   # Listfile containing dataset splits
   wget http://cnrpark.it/dataset/splits.zip
   unzip splits.zip
   

4. Add a config.py files inside each folder in splits/ to tell pyffe where the images are. The content of the files should be like this (adjust the root_dir attribute to the absolute path of the extracted datasets):

   config = dict(root_folder = '/path/to/dataset/dir/')
   

   This path will be prepended to each line in the list files defining the various splits.

5. Train and evaluate all the models by running:

   python main.py
   

   Modify main.py to select the experiments you want to reproduce. Run pklot.py if you want to train and evaluate our architecture on the PKLot splits only.

Citation

@article{amato2017deep,
  title={Deep learning for decentralized parking lot occupancy detection},
  author={Amato, Giuseppe and Carrara, Fabio and Falchi, Fabrizio and Gennaro, Claudio and Meghini, Carlo and Vairo, Claudio},
  journal={Expert Systems with Applications},
  volume={72},
  pages={327--334},
  year={2017},
  publisher={Pergamon}
}