An algorithm for automatic installation of detection area in CCTV image.

Auteur(s)
Lee, I.J. Min, J.Y. Choi, J.U. Namkoong, S. & Song, J.K.
Jaar
Samenvatting

In this paper, an algorithm is suggested for automatic instalment of detection area in CCTVs on highways. The automatic establishment of the detection area can provide to existing CCTV facilities extra functions that are typically imposed on image detectors. When edge detection method is used for automatic installation of detection area, it can provide relatively good results in daytime. However, edge-detection method cannot be applied in nighttime. For this reason, to provide reasonable solution to both daytime image and nighttime image, difference image method and morphology are applied to extract the contour line of roads. Then, Least Square Method (LSM) is employed to derive equations of lines indicating roadsides on the highway. The angle of CCTV camera facing the roads is estimated using the slope of the line's equations, and traffic lanes are derived by applying ratio of each lane in the roads. Finally, detection area is installed for each traffic lane in captured images. Experiments have been conducted on image data captured at CCTV cameras installed 15 meters high in Cheongju Interchange (IC) downstream, and camera installed at the entrance of Cheonggye tunnel along the Seoul metropolitan freeway. The equation of the lines of traffic lane is derived from the images of 640 x 480 pixel, 256 gray level, and detection areas are installed between the lanes. For the covering abstract see ITRD E114174.

Publicatie aanvragen

15 + 2 =
Los deze eenvoudige rekenoefening op en voer het resultaat in. Bijvoorbeeld: voor 1+3, voer 4 in.

Publicatie

Bibliotheeknummer
C 24596 (In: C 22454 CD-ROM) /73 / ITRD E115831
Uitgave

In: From vision to reality : proceedings of the 7th World Congress on Intelligent Transportation Systems ITS, Turin, Italy, 6-9 November 2000, 10 p., 4 ref.

Onze collectie

Deze publicatie behoort tot de overige publicaties die we naast de SWOV-publicaties in onze collectie hebben.