Design And Implementation Of Rehabilitation Detection And Training System

Along with social progress in recent years, monitoring technology is widely used in public areas, leading to a rapid increase in surveillance video, the traditional fast-forward, rewind the video query increasingly unable to meet the demand. Traditional query not only consume a lot of queries by time and energy, but also easy to miss the target information. Abstract technology makes video appears low status of video query efficiency has been greatly improved, however difficult the video summary technology into practical use is a summary of the main face of the problem of slow extraction rate. Being taken primarily by single extraction mode to speed up the performance of computer hardware upgrade summary extraction rate. This approach not only limited room for improvement, and costly, especially easy to reduce the useful life of equipment, resulting in a waste of resources.This article attempts to use cloud computing technology to make the current mainstream movement based on dynamic video summary of technical transformation into distributed by a stand-alone mode extraction mode, the face of data expansion, just to improve the video summary extraction rate by adding low-cost computing nodes, Abstract technology can make a better video into actual production.First, a summary of cloud computing technology and video technology for a lot of research. Select the appropriate open source cloud computing platform Hadoop distributed systems, distributed systems and Hadoop processing video summaries and related mechanisms were analyzed. In the video summary technology, an overview of the general situation of the video summary technology, detailing the main course of the campaign process is based on a summary motion video technology.Secondly, on the Hadoop distributed cloud application development system of rules and video processing technology characteristics were studied, pointed out the use of Hadoop distributed systems handle video summary of advantages, but made a video summary to be Mapreduce, frame integrity, Consistency sports events and other issues. Through the process of analysis, Hadoop data storage expansion process, increasing block logic, mathematical simulation of the sport events, taking reasonable threshold, etc., gives a solution, ready to build Hadoop-based video summarization system a theoretical basis.Again, based on the above theoretical analysis, the development of Hadoop-based video summarization system conducted requirements analysis, design the architecture of the whole system, to achieve the entire system.Finally, the design of the experimental program, the experimental data showed that when the amount of data increases to350MB, a two-node Hadoop-based video summarization system relative to the stand-alone mode extraction efficiency is higher. And with the increase of data, efficiency of Hadoop-based video summarization system is becoming more and eventually processing cycles tend to1.65in stand-alone mode, and has better scalability.