Video Quality Diagnostic Studies Based On The Broadband Communication Network

With the increasing development of broadband communication network transmission, a large number of popular video businesses put forward higher requirements for the application of video surveillance in the security monitoring industry and highway monitoring. The image video quality is the premise of the video surveillance. Only the clear images can reflect the real situation in the video surveillance area, so the video quality diagnostic research has became the focus of attention.The subject comes from the cooperation projects between our laboratory and Beijing VORX Communication Technology Co Ltd, I complete the video image quality diagnostics in the transmission network by the company’s existing network video transmission system. Working of this paper is listed as below:1. The essential content of the video image analysis is analyzing the characteristics of the image, therefore, I first intensive studies for the classification of the image feature and the different types of image feature extraction algorithm, realized image feature extraction such as the image color features, texture features, corner features, as well as image wavelet transform This is the basis of the image analysis, also is the basis of this article. Quickly and accurately achieving the distinction of image class and the diagnostic of the video quality was based on it.2. This paper has deeply studying the image comparison algorithm and machine learning algorithms, learning the open source library for image processing called opencv. this paper could judge whether the two video images are the same images fast and accurately, judge movement direction between the two near video images.Structuring learning environment of BP neural network with opencv.3. This paper completed the video diagnostics system, designed four types of BP neural network to detect video quality fault diagnosis including abnormal clarity, color cast, abnormal brightness and stripes noise of video signal, made2types of fault diagnosis for the disappeared video signal and the screen freezing video signal by analyzing whether the two video images are same, detected whether the rotation of PTZ was abnormal by judging the movement direction between the two near video images when PTZ was moving.I intercepted a large number of image data from the company’s existing urban monitoring network to test the functionality of video diagnostic system. The result showed that this video diagnostic system for the7types of video fault had a high degree of accuracy, and it had a very high value.