Á¤º¸°úÇÐȸ ÄÄÇ»ÆÃÀÇ ½ÇÁ¦ ³í¹®Áö (KIISE Transactions on Computing Practices)
Current Result Document :
ÇѱÛÁ¦¸ñ(Korean Title) |
µµ½Ã ±³Åë È帧 ¿¹ÃøÀ» À§ÇÑ CCTV ºñµð¿À ó¸® ¹× ±³Åë ³×Æ®¿öÅ© ¸ðµ¨¸µ ±â¼ú |
¿µ¹®Á¦¸ñ(English Title) |
CCTV Video Handling and Traffic Network Modeling Technique for Urban Traffic Flow Prediction |
ÀúÀÚ(Author) |
¿¬ÇѺ°
¼¼º¹ü
È«ÇýÀÎ
ˌ ˱
Hanbyul Yeon
Seongbum Seo
Hyein Hong
Yun Jang
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 27 NO. 06 PP. 0289 ~ 0294 (2021. 06) |
Çѱ۳»¿ë (Korean Abstract) |
º» ³í¹®Àº ¿©·¯ CCTV ºñµð¿À¸¦ »ç¿ëÇÏ¿© µµ½Ã ±³Åë È帧À» ¸ðµ¨¸µÇϱâ À§ÇÑ ÇÁ·¹ÀÓ¿öÅ©¸¦ Á¦½ÃÇÑ´Ù. ÇÁ·¹ÀÓ¿öÅ©´Â CCTV ¿µ»ó¿¡¼ ½Ç½Ã°£À¸·Î Â÷·®À» ŽÁöÇÏ¿© Â÷·® È帧 µ¥ÀÌÅ͸¦ »ý¼ºÇÑ´Ù. ±×·± ´ÙÀ½ Â÷·® È帧 µ¥ÀÌÅÍ¿Í CCTV Ä«¸Þ¶ó ³×Æ®¿öÅ©¸¦ °áÇÕÇÏ¿© µµ½Ã ±³Åë ³×Æ®¿öÅ© µ¥ÀÌÅ͸¦ »ý¼ºÇÑ´Ù. ´ëºÎºÐÀÇ CCTV Ä«¸Þ¶ó´Â ±³Â÷·ÎÀÇ ¸ðµç ¿µ¿ªÀ» ÃÔ¿µÇÏÁö ¾Ê´Â´Ù. ±×·¯¹Ç·Î ±³Â÷·Î¿Í ¿¬°áµÈ ¸ðµç µµ·Î¿¡ ´ëÇÑ ±³Åë È帧 µ¥ÀÌÅ͸¦ ¾òÀ» ¼ö ¾ø´Ù. ³í¹®¿¡¼´Â È®ÀåµÈ DCRNN ¸ðµ¨À» »ç¿ëÇÏ¿© ±³Â÷·Î¿Í ¿¬°áµÈ µµ·Î¿¡¼ °üÃøµÇÁö ¾ÊÀº Â÷·®ÀÇ ±³Åë È帧À» ÃßÁ¤ÇÑ´Ù
|
¿µ¹®³»¿ë (English Abstract) |
In this paper, we present a framework for modeling urban traffic networks with multiple Closed-Circuit Television(CCTV) videos. We extract vehicle flow data by detecting vehicles in realtime from the CCTV videos. We couple the vehicle flow data and the CCTV camera network to generate the urban traffic network data. Since most CCTV cameras do not record the whole intersection area, it is not possible to obtain the traffic flow data for every roadway connected to the intersection. We use the extended Diffusion Convolutional Recurrent Neural Network (DCRNN) to estimate the traffic flows on roadways that are not observed through the CCTV cameras.
|
Å°¿öµå(Keyword) |
±³Åë ³×Æ®¿öÅ© ¸ðµ¨¸µ
±³Åë È¥À⠺м®
È®»ê ÄÁº¼·ç¼Ç ¼øȯ ½Å°æ¸Á
DCRNN
traffic network modeling
traffic congestion analysis
diffusion convolutional recurrent neural network
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|