Deep Learning Based Parking Vacancy Detection for Smart Cities

Date

2022-01-04

Contributor

Advisor

Department

Instructor

Depositor

Speaker

Researcher

Consultant

Interviewer

Narrator

Transcriber

Annotator

Journal Title

Journal ISSN

Volume Title

Publisher

Volume

Number/Issue

Starting Page

Ending Page

Alternative Title

Abstract

Parking shortage is a major problem in modern cities. Drivers cruising in search of a parking space directly translate into frustration, traffic congestion, and excessive carbon emission. We introduce a simple and effective deep learning-based parking space notification (PSN) system to inform drivers of new parking availabilities and re-occupancy of the freed spaces. Our system is particularly designed to target areas with severe parking shortages (i.e., nearly all parking spaces are occupied), a situation that allows us to convert the problem of detecting parking vacancies into recognizing vehicles leaving from their stationary positions. Our PSN system capitalizes on a calibrated Mask R-CNN model and a unique adaptation of the IoU concept to track the changes of vehicle positions in a video stream. We evaluated PSN using videos from a CCTV camera installed at a private parking lot and publicly available YouTube videos. The PSN system successfully captured all new parking vacancies arising from leaving vehicles with no false positive detections. Prompt notification messages were sent to users via cloud messaging services.

Description

Keywords

Smart (City) and Data Streaming Application Development: Challenges and Experiences, deep learning, iot, object localization, parking vacancy detection, smart cities

Citation

Extent

7 pages

Format

Geographic Location

Time Period

Related To

Proceedings of the 55th Hawaii International Conference on System Sciences

Related To (URI)

Table of Contents

Rights

Attribution-NonCommercial-NoDerivatives 4.0 International

Rights Holder

Local Contexts

Email libraryada-l@lists.hawaii.edu if you need this content in ADA-compliant format.