DSpace Repository

Improving Traffic Law Enforcement Accuracy with Violating Vehicle Tracking and "Best Frame" Analysis

Show simple item record

dc.contributor.advisor Dailey, Matthew
dc.contributor.author Kommanaboina, Balakrishna Phani
dc.contributor.other Mongkol, Ekpanyapong
dc.contributor.other Attaphongse, Taparugssanagorn
dc.date.accessioned 2019-12-23T02:33:25Z
dc.date.available 2019-12-23T02:33:25Z
dc.date.issued 2019-12
dc.identifier.other AIT
dc.identifier.uri http://www.cs.ait.ac.th/xmlui/handle/123456789/954
dc.description 57 p. en_US
dc.description.abstract Traffic surveillance is one of the important measures that can reduce number of road accidents and collisions. Cameras can play an essential role in collision detection, traffic management, and traffic law enforcement. Applications using computer vision include vehicle tracking and speed estimation. Lane change violations are one type of moving violation in which a vehicle crosses a solid white line demarking the lane. In collecting evidence from a lane change violation, we may simply extract the frame in which the violation occurred but it would be better to extract the "best frame" in the video sequence. This method results in reducing the redundant and incorrect data frames. The dataset used in this study contains overview traffic footage of Phuket,Thailand. The system is trained for close traffic area with semi-densed traffic for Object detection and tracking. The main aim of this research study is the selection of "best frame"(containing license plate of the vehicle) for a violating track among the sequence of frames it has been detected in the video. The "best frame" is extracted on the basis of different criteria such as area, magnitude, sharpness of detected frames. The "best frame" is extracted by taking the maximum of product of area and sharpness, where the magnitude is measured on the basis of Frobenius norm and sharpness value is measured on the basis of Laplace operator. A list of comparison with violating frame and "best frame" are shown in the results. With the help of Vinfo, a real time surveillance system for vehicle license plates' Optical Character Recognition the analysis of this method is measured, which shows "best frame" for every image except for fragmented tracks. en_US
dc.description.sponsorship AIT Fellowship en_US
dc.language.iso en en_US
dc.publisher AIT en_US
dc.subject Object Detection en_US
dc.subject Object Tracking en_US
dc.subject Frobenius norm en_US
dc.subject Laplace Operator en_US
dc.subject Optical Character Recognition. en_US
dc.subject Vinfo en_US
dc.title Improving Traffic Law Enforcement Accuracy with Violating Vehicle Tracking and "Best Frame" Analysis en_US
dc.type Research report en_US


Files in this item

This item appears in the following Collection(s)

Show simple item record

Search DSpace


Advanced Search

Browse

My Account