Anes Madani, Suman Kumar, Linh Ba Nguyen, Jiling Zhong
{"title":"A Robust Road Region of Interest Identification Scheme for Traffic-Video Data Mining","authors":"Anes Madani, Suman Kumar, Linh Ba Nguyen, Jiling Zhong","doi":"10.1109/ICCNC.2019.8685513","DOIUrl":null,"url":null,"abstract":"Traffic video data mining applications demand a road region of interest be identified. Typically, the region of interest is drawn manually, thus making it challenging to design large scale data mining applications utilizing widely available open access live stream traffic cameras since diverse scenarios require diverse region drawings. This paper presents a novel algorithm to identify road region of interest, therefore, automating the otherwise a manual process and making it applicable to diverse traffic live stream scenarios encountered in practice. The algorithm utilizes problem domain property of vehicle mobility constraints. Through experimentation, we show that algorithm is robustly resistant to the wide variety of cases of camera resolution, traffic volume, light condition, camera shakiness etc. The algorithm aims to simplify the overall design of large scale open camera traffic video mining task to aid next generation transportation-data-as-a-service based applications.","PeriodicalId":161815,"journal":{"name":"2019 International Conference on Computing, Networking and Communications (ICNC)","volume":"28 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2019-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2019 International Conference on Computing, Networking and Communications (ICNC)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICCNC.2019.8685513","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
Traffic video data mining applications demand a road region of interest be identified. Typically, the region of interest is drawn manually, thus making it challenging to design large scale data mining applications utilizing widely available open access live stream traffic cameras since diverse scenarios require diverse region drawings. This paper presents a novel algorithm to identify road region of interest, therefore, automating the otherwise a manual process and making it applicable to diverse traffic live stream scenarios encountered in practice. The algorithm utilizes problem domain property of vehicle mobility constraints. Through experimentation, we show that algorithm is robustly resistant to the wide variety of cases of camera resolution, traffic volume, light condition, camera shakiness etc. The algorithm aims to simplify the overall design of large scale open camera traffic video mining task to aid next generation transportation-data-as-a-service based applications.