Automatic Traffic Surveillance System for Traffic Conjesion
碩士 === 國防大學中正理工學院 === 資訊科學研究所 === 96 === This paper presents an automatic traffic flow and speed surveillance system from traffic jam video sequences. We use only one camera to get the sequence of traffic images, and the camera is mounted on top of the road. The system uses moving object detection t...
Main Authors: | , |
---|---|
Other Authors: | |
Format: | Others |
Language: | zh-TW |
Published: |
2008
|
Online Access: | http://ndltd.ncl.edu.tw/handle/44601363714121659540 |
id |
ndltd-TW-096CCIT0394015 |
---|---|
record_format |
oai_dc |
spelling |
ndltd-TW-096CCIT03940152016-05-16T04:09:41Z http://ndltd.ncl.edu.tw/handle/44601363714121659540 Automatic Traffic Surveillance System for Traffic Conjesion 壅塞道路之影像式自動車流與車速偵測系統 Liu, shih-cheng 劉世程 碩士 國防大學中正理工學院 資訊科學研究所 96 This paper presents an automatic traffic flow and speed surveillance system from traffic jam video sequences. We use only one camera to get the sequence of traffic images, and the camera is mounted on top of the road. The system uses moving object detection to detect vehicles, and then to classify and count the vehicles which are segmented by the proposed recognition and tracking methods. Because the camera is mounted on top of the road and has a depression angle, we also consider the problems of the vehicle occlusion. When two or more vehicles are too close, they will be segmented by an occlusion segmentation algorithm. Therefore, this paper proposes an occlusion segmentation method to solve the occlusion problem. When the vehicles are occlusive, the system will segment continuously the merging vehicles in the image sequence, and then keeping recognize the vehicles after segmented and track the vehicles until they leave the image frame. Different from traditional methods that can only divide two occlusive vehicles, the proposed method can segment the vehicles from traffic jam and track each occlusive vehicle. The system will segment and recognize the occlusive vehicles before they go into the traffic jam area, and it still can recognize and track the vehicles after they go into the traffic jam area. In this paper we use many video sequences of traffic jam images to verify the proposed system. From these experiments, the system can segment and track the occlusive vehicles from the occlusion and traffic jam issues. In the future work, we hope to improve the detection rate of occlusion segmentation and enhance the efficiency for tracking the occlusive vehicles from different traffic jam conditions. The system has been continuously monitoring traffic conditions on city roads and highways for 24 hours without any alteration of parameters. 羅裕群、瞿忠正 2008 學位論文 ; thesis 54 zh-TW |
collection |
NDLTD |
language |
zh-TW |
format |
Others
|
sources |
NDLTD |
description |
碩士 === 國防大學中正理工學院 === 資訊科學研究所 === 96 === This paper presents an automatic traffic flow and speed surveillance system from traffic jam video sequences. We use only one camera to get the sequence of traffic images, and the camera is mounted on top of the road. The system uses moving object detection to detect vehicles, and then to classify and count the vehicles which are segmented by the proposed recognition and tracking methods. Because the camera is mounted on top of the road and has a depression angle, we also consider the problems of the vehicle occlusion. When two or more vehicles are too close, they will be segmented by an occlusion segmentation algorithm. Therefore, this paper proposes an occlusion segmentation method to solve the occlusion problem. When the vehicles are occlusive, the system will segment continuously the merging vehicles in the image sequence, and then keeping recognize the vehicles after segmented and track the vehicles until they leave the image frame. Different from traditional methods that can only divide two occlusive vehicles, the proposed method can segment the vehicles from traffic jam and track each occlusive vehicle. The system will segment and recognize the occlusive vehicles before they go into the traffic jam area, and it still can recognize and track the vehicles after they go into the traffic jam area. In this paper we use many video sequences of traffic jam images to verify the proposed system. From these experiments, the system can segment and track the occlusive vehicles from the occlusion and traffic jam issues.
In the future work, we hope to improve the detection rate of occlusion segmentation and enhance the efficiency for tracking the occlusive vehicles from different traffic jam conditions. The system has been continuously monitoring traffic conditions on city roads and highways for 24 hours without any alteration of parameters.
|
author2 |
羅裕群、瞿忠正 |
author_facet |
羅裕群、瞿忠正 Liu, shih-cheng 劉世程 |
author |
Liu, shih-cheng 劉世程 |
spellingShingle |
Liu, shih-cheng 劉世程 Automatic Traffic Surveillance System for Traffic Conjesion |
author_sort |
Liu, shih-cheng |
title |
Automatic Traffic Surveillance System for Traffic Conjesion |
title_short |
Automatic Traffic Surveillance System for Traffic Conjesion |
title_full |
Automatic Traffic Surveillance System for Traffic Conjesion |
title_fullStr |
Automatic Traffic Surveillance System for Traffic Conjesion |
title_full_unstemmed |
Automatic Traffic Surveillance System for Traffic Conjesion |
title_sort |
automatic traffic surveillance system for traffic conjesion |
publishDate |
2008 |
url |
http://ndltd.ncl.edu.tw/handle/44601363714121659540 |
work_keys_str_mv |
AT liushihcheng automatictrafficsurveillancesystemfortrafficconjesion AT liúshìchéng automatictrafficsurveillancesystemfortrafficconjesion AT liushihcheng yōngsāidàolùzhīyǐngxiàngshìzìdòngchēliúyǔchēsùzhēncèxìtǒng AT liúshìchéng yōngsāidàolùzhīyǐngxiàngshìzìdòngchēliúyǔchēsùzhēncèxìtǒng |
_version_ |
1718268030467702784 |