Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Using Deep Neural Networks to Quantify Parking Dwell Time

Marcelo Eduardo Marques Ribas, Heloisa Benedet Mendes, Luiz Eduardo Soares de Oliveira, Luiz Antonio Zanlorensi, Paulo Ricardo Lisboa de Almeida

TL;DR

This work proposes a method that combines two deep neural networks to compute the dwell time of each car in a parking lot using a Siamese network and shows that the proposed Siamese network is promising but impacted by the quality of the classifier used at the beginning of the pipeline.

Abstract

In smart cities, it is common practice to define a maximum length of stay for a given parking space to increase the space's rotativity and discourage the usage of individual transportation solutions. However, automatically determining individual car dwell times from images faces challenges, such as images collected from low-resolution cameras, lighting variations, and weather effects. In this work, we propose a method that combines two deep neural networks to compute the dwell time of each car in a parking lot. The proposed method first defines the parking space status between occupied and empty using a deep classification network. Then, it uses a Siamese network to check if the parked car is the same as the previous image. Using an experimental protocol that focuses on a cross-dataset scenario, we show that if a perfect classifier is used, the proposed system generates 75% of perfect dwell time predictions, where the predicted value matched exactly the time the car stayed parked. Nevertheless, our experiments show a drop in prediction quality when a real-world classifier is used to predict the parking space statuses, reaching 49% of perfect predictions, showing that the proposed Siamese network is promising but impacted by the quality of the classifier used at the beginning of the pipeline.

Using Deep Neural Networks to Quantify Parking Dwell Time

TL;DR

This work proposes a method that combines two deep neural networks to compute the dwell time of each car in a parking lot using a Siamese network and shows that the proposed Siamese network is promising but impacted by the quality of the classifier used at the beginning of the pipeline.

Abstract

In smart cities, it is common practice to define a maximum length of stay for a given parking space to increase the space's rotativity and discourage the usage of individual transportation solutions. However, automatically determining individual car dwell times from images faces challenges, such as images collected from low-resolution cameras, lighting variations, and weather effects. In this work, we propose a method that combines two deep neural networks to compute the dwell time of each car in a parking lot. The proposed method first defines the parking space status between occupied and empty using a deep classification network. Then, it uses a Siamese network to check if the parked car is the same as the previous image. Using an experimental protocol that focuses on a cross-dataset scenario, we show that if a perfect classifier is used, the proposed system generates 75% of perfect dwell time predictions, where the predicted value matched exactly the time the car stayed parked. Nevertheless, our experiments show a drop in prediction quality when a real-world classifier is used to predict the parking space statuses, reaching 49% of perfect predictions, showing that the proposed Siamese network is promising but impacted by the quality of the classifier used at the beginning of the pipeline.

Paper Structure

This paper contains 11 sections, 2 equations, 6 figures, 5 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Works
  3. Proposed Approach
  4. Experimental Protocol
  5. Datasets and Metrics
  6. Networks and Training Procedure
  7. Experiments
  8. Measuring the Dwell Time using the Ground Truth
  9. Measuring the Dwell Time using a Classifier
  10. Lessons Learned
  11. Conclusion

Figures (6)

  • Figure 1: A car parked in the same position at different times -- PKLot dataset.
  • Figure 2: Image examples from the PKLot and CNRPark-EXT datasets.
  • Figure 3: Siamese network training samples examples.
  • Figure 4: Flowchart example for two parking spaces. Both parking spaces were occupied between $t-k$ and $t$. The parked car in the orange parking space is not the same between images, and a new timer is created. The car in the blue parking space remained the same, and its timer was increased.
  • Figure 5: Car's Dwell Time Histogram using the a) PKLot and b) CNRPark-EXT datasets as test sets.
  • ...and 1 more figures