File Download
  Links for fulltext
     (May Require Subscription)
Supplementary

Article: Turning traffic surveillance cameras into intelligent sensors for traffic density estimation

TitleTurning traffic surveillance cameras into intelligent sensors for traffic density estimation
Authors
KeywordsCamera calibration
Traffic density estimation
Traffic surveillance camera
Vehicle detection
Issue Date22-Jun-2023
PublisherSpringer
Citation
Complex & Intelligent Systems, 2023, v. 9, n. 6, p. 7171-7195 How to Cite?
Abstract

Accurate traffic density plays a pivotal role in the Intelligent Transportation Systems (ITS). The current practice to obtain the traffic density is through specialized sensors. However, those sensors are placed in limited locations due to the cost of installation and maintenance. In most metropolitan areas, traffic surveillance cameras are widespread in road networks, and they are the potential data sources for estimating traffic density in the whole city. Unfortunately, such an application is challenging since surveillance cameras are affected by the 4L characteristics: Low frame rate, Low resolution, Lack of annotated data, and Located in complex road environments. To the best of our knowledge, there is a lack of holistic frameworks for estimating traffic density from traffic surveillance camera data with 4 L characteristics. Therefore, we propose a framework for estimating traffic density using uncalibrated traffic surveillance cameras. The proposed framework consists of two major components: camera calibration and vehicle detection. The camera calibration method estimates the actual length between pixels in the images and videos, and the vehicle counts are extracted from the deep-learning-based vehicle detection method. Combining the two components, high-granular traffic density can be estimated. To validate the proposed framework, two case studies were conducted in Hong Kong and Sacramento. The results show that the Mean Absolute Error (MAE) for the estimated traffic density is 9.04 veh/km/lane in Hong Kong and 7.03 veh/km/lane in Sacramento. The research outcomes can provide accurate traffic density without installing additional sensors.


Persistent Identifierhttp://hdl.handle.net/10722/339090
ISSN
2023 Impact Factor: 5.0
2023 SCImago Journal Rankings: 1.321
ISI Accession Number ID

 

DC FieldValueLanguage
dc.contributor.authorHu, ZJ-
dc.contributor.authorLam, WHK-
dc.contributor.authorWong, SC-
dc.contributor.authorChow, AHF-
dc.contributor.authorMa, W-
dc.date.accessioned2024-03-11T10:33:49Z-
dc.date.available2024-03-11T10:33:49Z-
dc.date.issued2023-06-22-
dc.identifier.citationComplex & Intelligent Systems, 2023, v. 9, n. 6, p. 7171-7195-
dc.identifier.issn2199-4536-
dc.identifier.urihttp://hdl.handle.net/10722/339090-
dc.description.abstract<p>Accurate traffic density plays a pivotal role in the Intelligent Transportation Systems (ITS). The current practice to obtain the traffic density is through specialized sensors. However, those sensors are placed in limited locations due to the cost of installation and maintenance. In most metropolitan areas, traffic surveillance cameras are widespread in road networks, and they are the potential data sources for estimating traffic density in the whole city. Unfortunately, such an application is challenging since surveillance cameras are affected by the <strong>4L</strong> characteristics: <strong>L</strong>ow frame rate, <strong>L</strong>ow resolution, <strong>L</strong>ack of annotated data, and <strong>L</strong>ocated in complex road environments. To the best of our knowledge, there is a lack of holistic frameworks for estimating traffic density from traffic surveillance camera data with 4 L characteristics. Therefore, we propose a framework for estimating traffic density using uncalibrated traffic surveillance cameras. The proposed framework consists of two major components: camera calibration and vehicle detection. The camera calibration method estimates the actual length between pixels in the images and videos, and the vehicle counts are extracted from the deep-learning-based vehicle detection method. Combining the two components, high-granular traffic density can be estimated. To validate the proposed framework, two case studies were conducted in Hong Kong and Sacramento. The results show that the Mean Absolute Error (MAE) for the estimated traffic density is 9.04 veh/km/lane in Hong Kong and 7.03 veh/km/lane in Sacramento. The research outcomes can provide accurate traffic density without installing additional sensors.</p>-
dc.languageeng-
dc.publisherSpringer-
dc.relation.ispartofComplex & Intelligent Systems-
dc.rightsThis work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.-
dc.subjectCamera calibration-
dc.subjectTraffic density estimation-
dc.subjectTraffic surveillance camera-
dc.subjectVehicle detection-
dc.titleTurning traffic surveillance cameras into intelligent sensors for traffic density estimation-
dc.typeArticle-
dc.description.naturepublished_or_final_version-
dc.identifier.doi10.1007/s40747-023-01117-0-
dc.identifier.scopuseid_2-s2.0-85163006172-
dc.identifier.volume9-
dc.identifier.issue6-
dc.identifier.spage7171-
dc.identifier.epage7195-
dc.identifier.eissn2198-6053-
dc.identifier.isiWOS:001014507600001-
dc.identifier.issnl2199-4536-

Export via OAI-PMH Interface in XML Formats


OR


Export to Other Non-XML Formats