A systematic review on computer vision-based parking lot management applied on public datasets
作者：
Highlights：
• Robust publicly available parking lot images datasets are reviewed and compared.
• Most works present biased results and lack standard protocols.
• Open issues like dataset dependency and parking spot detection are debated.
• New datasets containing diverse luminosity, climates and ground-truths are needed.
• New test protocols are needed, including standard metrics and cross-dataset tests.
摘要
•Robust publicly available parking lot images datasets are reviewed and compared.•Most works present biased results and lack standard protocols.•Open issues like dataset dependency and parking spot detection are debated.•New datasets containing diverse luminosity, climates and ground-truths are needed.•New test protocols are needed, including standard metrics and cross-dataset tests.
论文关键词：Parking lot,Dataset,Benchmark,Machine learning,Image processing
论文评审过程：Received 2 July 2021, Revised 21 September 2021, Accepted 22 February 2022, Available online 12 March 2022, Version of Record 25 March 2022.
论文官网地址：https://doi.org/10.1016/j.eswa.2022.116731
