System Design Of Real-time RGB Data Comparison TV Symbol Recognition Based On Matlab

Besides as exclusive labels for different TV signals, TV station logos can be as a main reflection of TV signals being transmitted normally and not being inter-cut or being closed down. TV signals are packed in generator rooms of TV stations and transmitted to satellite earth stations by exclusive optical fiber or micro-wave equipment. They are sent to antenna system by coding and multiplexing and then sent to satellite transponders in radio frequency. This process is for real-time transmission. Due to the existence of interference, especially the fact that organizations had repeatedly attempted to inter-cut the illegal signal frequency, it turns to be important to closely monitor transmission of signals.If automatic and intelligent monitoring judgment to simultaneously transmitted and received television signals by computer technology can be used, it will greatly reduce the delay of manual operation and improve the response speed of failure recovery so as to avoid broadcast accidents. Rapid and accurate recognition of illegal accidental pictures is also the judgment basis for operations such as active switchover of multi-channel signal source by intelligent switching system.This paper puts forward a method of real-time identification and alignment to TV video logo using digital image processing technique. Firstly, using the space-time invariance features of TV logos, a number of points at the regions where TV signals regularly appear are deployed, which will form an exclusive lattice of some channel logo. All lattice information of every different channel logo exists in specifically set database respectively. Secondly, the algorithm is used to read RGB three-color information of corresponding pixel of each frame image lattice. And then, weighted differences of real-time received signal and a reference signal RGB information are compared. If the weighted difference is greater than the threshold value, a warning is issued and the relevant standby video clips for daily check are recorded. Considering the different weights of importance and key value of different regions, so weighting factors are appropriately applied to the difference calculation formula to distinguish the degrees of importance and priority of different points, in order to reduce judgment accuracy effect on the logo's transparent or less important regions, and reduce false positives.Considering the high requirements raised by satellite earth stations to the real-time nature and reliability of monitoring, this paper proposes the algorithm of real-time extraction and matching of RGB information of low-level features. By avoiding complex algorithms of logo template extraction, storage and identifying training, it can ensure that the computer system resource is of low occupancy and the hardware is of relatively low.