| With the promotion of digital television, Effective monitoring of digital televisionbroadcast plays a more and more important role in guaranteeing the quality and safety ofbroadcast. The automatically video contents monitoring, using computer imageidentification technology, is an important part of monitoring system of television broadcast.The algorithms of frozen-frame and black-frame detection in TrinityAres monitoringsystem of digital television broadcast of BoHui Technology Co, Ltd. are introduced.Blocking-error frame detection is a blank in civil video monitoring systems of digitaltelevision broadcast. This paper focuses on the detection of blocking-error frame. Thecauses and characters of blocking-error frame are analyzed in the view of bit stream erroranalysis. The uses of classical edge detection algorithms are analyzed. A new algorithm forblocking-error frame detection, using difference edge detection, is researched andprogrammed in the TrinityAres system, in VC++. Experiments prove that blocking-errorframes can be efficiently detected.Firstly, the system part, compression coding and key technology of MPEG-2areintroduced. After this, the algorithms of image identification are researched such aspretreatment, segmentation, edge detection and feature extraction. Secondly, the charactersand the detection of black-frame and frozen-frame are introduced in background of theTrinityAres monitoring system of digital television broadcast. Lastly, the characters ofblocking-error frame are analyzed, including the blocking-error caused by bit errors ofdifferent locations, the transmission of blocking-error in space and time. A new algorithmfor blocking-error frame detection, using difference edge of gray images and the featuresof U, V signals, is researched and programmed in VC++in TrinityAres monitoring Systemof digital television broadcast. The result of a series of experiments shows that the aim ofdetecting blocking-error frame is achieved with superior accuracy. |
PDF Full Text Request