Research And Implementation Of Video Scene Detection Technology In Broadband Digital Trunking Terminals

The broadband digital trunking system is a professional command and dispatch system that integrates various services such as voice,data,image and video.It is mainly used in government,public security,emergency rescue,logistics,and military departments.Video service in broadband digital trunking terminals has the function of real-time video transmission and is an important business of the broadband digital trunking system.In order to meet the high reliability and low latency requirements of the trunking communication system,broadband digital trunking terminals need to transmit high-quality and real-time video.Video scene detection technology can extract real-time video scenes,and dynamically adjust the bit rate and resolution parameters of real-time video according to the scene,ensuring the quality of real-time video in different network environments.Traditional scene detection technology cannot meet the broadband digital trunking terminals’ requirements for high accuracy and low latency.Therefore,the research and implementation of video scene detection technology in broadband digital trunking terminals is of great significance.On the basis of reviewing the research status of broadband digital trunking system and scene detection technology at home and abroad,this paper studies and realizes the scene detection technology in broadband digital trunking terminals.The main work is as follows:1.A software architecture of broadband digital trunking terminals based on scene detection technology is proposed.This architecture realizes the video conferencing function of broadband digital trunking terminals and video scene detection technology.It has the characteristics of strong scalability,fast signaling transmission and reception,and easy maintenance,which solves the problem that the original broadband digital trunking terminals do not support scene detection.2.A weighted histogram detection algorithm based on adaptive threshold named WHDAAT is proposed in this paper.The algorithm comprehensively considers the grayscale histogram characteristics and brightness information of the video.By improving the gray-scale histogram algorithm and using adaptive thresholds,the accuracy of scene detection is effectively improved while ensuring the delay,and achieve a balance between accuracy and delay.3.Using Java and C programming languages,based on the Android operating system,the broadband digital trunking terminals video conferencing module,video transmission module and video scene detection module was developed.Realized the video scene detection technology in broadband digital trunking terminals.4.A video quality evaluation platform dedicated to broadband digital trunking terminals has been built,explains the verification scheme and verification content of the video scene detection technology in the broadband digital trunking terminals,verifies the video conferencing function,verifies the effectiveness of the scene detection technology,and analyzes the accuracy,recall,harmonic mean,and average delay performance of scene detection technology.The verification results show that the implemented video scene detection technology in broadband digital trunking terminals meet the requirements of accuracy and time delay.