Advertising Monit System Based On Fpga In The Research And Implementation

Advertisement detecting is mainly applied to collect delivery statistics of the sametrade. By doing this, the company can define their own advertising content and deliverymode. Further more, they can analyze the media’s role in product promotion. At thesame time, the media can be provided the basis for the performance of the contract byusing this.Advertisement detecting system is divided into the part of audio monitoring andthe part of TV station caption identification. These two parts operate independently. Thecore modules are the feature extraction and template matching.Comparing a variety of algorithms audio feature extraction, the sliding windowaccumulation algorithm which is not sensitive to the starting position with small amountof data is designed for audio monitoring. What’s more, the user can add thecharacteristics of the database by just using one key. Six channel audio monitoringworks in parallel in one system that can reduce the expense. Model-Based Design isused for audio Filtering. The HDL code is automatically generated by filterbuilder.Both the algorithm of dynamic cancellation and the algorithm of thresholdsegmentation with the sliding window template are used for the feature extraction of TVstation caption identifying. The recognition rate will not reduce with the movement ofthe TV station caption, which reflects the excellent anti-jitter and robustness. Furthermore, the database characteristics support updates.The template matching completes the intercomparison by searching for thecharacteristics consideration to time criterion and space criterion. Accordingly, themissing report rate and the false alert rate are effectively reduced.The simulation and cascade debugging is done for each module of the system. It isproved by the experiment that the system is stable and reliable.Advertisement detecting system based on FPGA which is composed in this thesishas Multi-functional features such as Intelligent Detection for advertisement detecting,establishing characteristics of the database in real time, identifying the caption of theTV station where the advertisement is broadcasting, etc. On the basis of the analysis, we may safely arrive at the conclusion that the advertisement detecting can be provided asufficient basis by using the system.