du.sePublications
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • chicago-author-date
  • chicago-note-bibliography
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Deep learning-based vehicle occupancy detection in an open parking lot using thermal camera
Dalarna University, School of Technology and Business Studies, Microdata Analysis.ORCID iD: 0000-0002-2078-3327
Dalarna University, School of Technology and Business Studies, Computer Engineering.ORCID iD: 0000-0002-1429-2345
Dalarna University, School of Technology and Business Studies, Information Systems.ORCID iD: 0000-0003-4812-4988
2019 (English)In: Article in journal (Other academic) Submitted
Abstract [en]

Parking vehicle is a daunting task and a common problem in many cities around the globe. The search for parking space leads to congestion, frustration and increased air pollution. Information of a vacant parking space would facilitate to reduce congestion and subsequent air pollution. Therefore, aim of the paper is to acquire vehicle occupancy in an open parking lot using deep learning. Thermal camera was used to collect the data during varying environmental conditions such as; sunny, dusk, dawn, dark and snowy conditions. Vehicle detection with deep learning was implemented where image classification and object localization were performed for multi object detection. The dataset consists of 527 images which were manually labelled as there were no pre-labelled thermal images available. Multiple deep learning networks such as Yolo, ReNet18, ResNet50 and GoogleNet with varying layers and architectures were evaluated on vehicle detection. Yolo, GoogleNet and ResNet18 are computationally efficient detectors which took less processing time while Resnet50 produced better detection results compared to other detectors. However, ResNet18 also produced minimal miss rates and is suitable for real time vehicle detection. The detected results were compared with a template of parking spaces and IoU value is used to identify vehicle occupancy information.

Place, publisher, year, edition, pages
2019.
National Category
Computer Vision and Robotics (Autonomous Systems)
Research subject
Complex Systems – Microdata Analysis
Identifiers
URN: urn:nbn:se:du-30605OAI: oai:DiVA.org:du-30605DiVA, id: diva2:1342141
Available from: 2019-08-12 Created: 2019-08-12 Last updated: 2019-08-14Bibliographically approved
In thesis
1. Developing decision support systems for last mile transportation problems
Open this publication in new window or tab >>Developing decision support systems for last mile transportation problems
2019 (English)Licentiate thesis, comprehensive summary (Other academic)
Abstract [en]

Last mile transportation is the most problematic phase of transportation needing additional research and effort. Longer waits or search times, lack of navigational directions and real-time information are some of the common problems associated with last mile transportation. Inefficient last mile transportation has an impact on the environment, fuel consumption, user satisfaction and business opportunities. Last mile problems exist in several transportation domains, such as: the landing of airplanes, docking of ships, parking of vehicles, attended home deliveries, etc. While there are dedicated inter-connected decision support systems available for ships and aircraft, similar systems are not widely utilized in parking or attended handover domains. Therefore, the scope of this thesis covers last mile transportation problems in parking and attended handover domains. One problem area for parking and attended handovers is due to lack of real-time information to the driver or consumer. The second problem area is dynamic scheduling where the handover vehicle must traverse additional distance to multiple handover locations due to lack of optimized routes. Similarly, during parking, lack of navigational directions to an empty parking space can lead to increased fuel consumption and CO2 emissions. Therefore, aim of this thesis is to design and develop decision support systems for last mile transportation problems by holistically addressing real time customer communication and dynamic scheduling problem areas. The problem areas discussed in this thesis consists of persistent issues even though they were widely discussed in the literature. In order to investigate the problem areas, microdata analysis approach was implemented in the thesis. The phases involved in Microdata analysis are: data collection, data processing, data storage, data analysis and decision-making. Other similar research domains, such as: computer science or statistics also involve phases such as data collection, processing, storage and analysis. These research domains also work in the fields of decision support systems or knowledge creation. However, knowledge creation or decision support systems is not a mandatory phase in these research domains, unlike Microdata analysis. Three papers are presented in this thesis, with two papers focusing on parking domains, while the third paper focuses on attended handover domains.

The first paper identifies available smart parking tools, applications and discusses their uses and drawbacks in relation to open parking lots. The usage of cameras in identifying parking occupancy was recognized as one of the suitable tools in this paper. The second paper uses a thermal camera to collect the parking lot data, while deep learning methodologies were used to identify parking occupancy detection. Multiple deep learning networks were evaluated for identifying parking spaces and one method was considered suitable for acquiring real time parking occupancy. The acquired parking occupancy information can be communicated to the user to address real-time customer communication problems. However, the decision support system (DSS) to communicate parking occupancy information still needs to be developed. The third paper focuses on the attended handovers domain where a decision support system was reported which addresses real-time customer communication and dynamic scheduling problems holistically. Based on a survey, customers accepted the use of mobile devices for enabling a real-time information flow for improving customer satisfaction. A pilot test on vehicle routing was performed where the decision support system reduced the vehicle routing distance compared to the route taken by the driver. The three papers work in developing decision support systems for addressing major last mile transportation problems in parking and attended handover domains, thus improving customer satisfaction, and business opportunities, and reducing fuel costs, and pollution.

Place, publisher, year, edition, pages
Borlänge: Dalarna University, 2019
Series
Dalarna Licentiate Theses in Microdata Analysis ; 10
Keywords
parking spaces, attended handovers, user satisfaction, pollution, business opportunities
National Category
Transport Systems and Logistics
Research subject
Complex Systems – Microdata Analysis
Identifiers
urn:nbn:se:du-30609 (URN)978-91-88679-02-4 (ISBN)
Presentation
2019-09-06, Clas Ohlsson, Borlänge, 13:00 (English)
Opponent
Supervisors
Available from: 2019-08-20 Created: 2019-08-14 Last updated: 2019-08-20

Open Access in DiVA

The full text will be freely available from 2019-10-31 21:26
Available from 2019-10-31 21:26

Authority records BETA

Paidi, VijayFleyeh, HasanNyberg, Roger G.

Search in DiVA

By author/editor
Paidi, VijayFleyeh, HasanNyberg, Roger G.
By organisation
Microdata AnalysisComputer EngineeringInformation Systems
Computer Vision and Robotics (Autonomous Systems)

Search outside of DiVA

GoogleGoogle Scholar

urn-nbn

Altmetric score

urn-nbn
Total: 52 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • chicago-author-date
  • chicago-note-bibliography
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf