Design And Implementation Of Video Surveillance System Based On Android

The contemporary society has gradually entered the era of mobile Internet.With the rapid development of wireless communication technology and improvement of the performance of mobile devices,traditional PC based video surveillance systems are moving to the mobile platform.The functions of mobile video surveillance systems in the present market are comparatively simple,for many applications can only play video without user friendly interface.To meet the user's diverse requirements,map services and IoTs are integrated into video surveillance in this thesis,which can provide more convenient functions and assist people to deal with emergencies.This thesis contributes to the design and implementation of a complete mobile video surveillance system.The server build with Spring MVC can send real-time messages to clients through MQTT protocol.The client is developed based on the Android platform,containing several modules such as video surveillance,alarm messages,IoT,map,login and settings.The followings are core modules of system:(1)Video surveillance module performs real-time preview,video playback and pan-tiltzoom control of the cameras.This thesis adopts a multi-thread player model to mitigate the effects of network jitter through a ring buffer.Considering both decoding performance and compatibility,MediaCodec class is used by default for hardware decoding and the FFmpeg media library is called for software decoding when MediaCodec cannot work.(2)Alarm message module is used to query and receive the alarm information from sensors,and the IoT module lists these sensors with current status.Both of the two modules receive messages from server through Android service components.A notification message will be generated when a new alarm message arrives.(3)Map is a common module for displaying the cameras and sensors in the map,which enables user to identify each device conveniently.On the other hand,this module provides navigation function which plays a auxiliary role in stuff scheduling in an emergency.Based on the above functions,this thesis optimizes the client from two perspectives.Firstly,a two-level cache mechanism of the camera thumbnail is designed.The remote images are cached into the memory and the external storage.It is verified by the test that the two-level cache can significantly increase the loading speed of the thumbnails.Secondly,the loading procedure of some Fragments is optimized to achieve the purpose of loading on demands,which can improve the performance of the client and reduce the pressure on the server.At the end of thesis,we perform the function test,performance test and compatibility test on the client.The results show that the client have a good performance in functionality,robustness and compatibility and can satisfy the demand of practical using.