Television station caption detection method based on edge detection

Abstract

The invention provides a television station caption detection method based on edge detection. The method comprises a data training module, an image acquisition module, an image preprocessing module and a station caption recognition module. The method comprises the following steps: acquiring station caption pictures of each television station; performing gray processing and edge detection processing on the station caption picture in the data set; constructing a convolutional neural network, and obtaining a station caption detection model; and issuing the station caption detection model to a station caption detection front end, inputting the real-time frame picture of the video into the model, and returning to judge whether the real-time frame picture contains the station caption and the type of the station caption. According to the method, gray processing and edge detection processing are carried out on the data set pictures, the original data volume of the images is reduced on the premise that station caption picture features are reserved, picture pixel points are converted into one-dimensional vectors, and the calculation amount in the data training process and the recognition process is effectively reduced. According to the invention, the efficiency of the station caption real-time detection process can be effectively improved.

Description

technical field [0001] The invention relates to the technical field of digital broadcasting and television, in particular to a method and device for detecting a station logo of a television station based on edge detection in a video image. Background technique [0002] With the majestic development of TV technology, my country's TV system has become a huge and complex system. During the front-end processing and network transmission of TV signals, transmission errors, artificial changes, and malicious attacks may occur. In order to ensure For the normal broadcasting of digital TV programs, the supervision of TV signals has become particularly important, and the detection of station logos is an important way to supervise TV signals. [0003] The traditional method of manually monitoring TV signals in real time is too powerful for the station logo detection method, and the detection efficiency is low, and there are also problems such as false negatives and other detection errors...

AI Technical Summary

Problems solved by technology

[0004] In order to solve the problems of low detection efficiency and high error rate of the above-mentioned station logo, the present invention provides a station logo detection method with fast recognition speed and high accuracy

Images