基于云存储技术的手语查询系统研究

发布时间：2018-09-05 16:48

【摘要】：随着社会和全球一体化的不断发展,用于人们交流的语言体系越来越丰富,而主要面向听障人交流的手语体系也相应的越来越庞大和繁杂。对手语进行说明描述和的收图片也越来越多,体积也将越来越大。为此,对手语进行描述的图片也越来越多,如何手语体系中海量的图片进行存储和查询管理也成为了信息时代手语发展的难点之一。本文主要采用基于Hadoop的云存储技术和基于Android的移动终端技术,为用户提供一个随时随地都能够快速查询的手语图示查询系统,这对于手语的应用和发展都具有非常重要的意义。论文对基于云存储技术的手语查询系统的研究主要包括如下几个方面的内容。(1)采用云技术提高手语查询性能本文采用云存储技术,当用户向服务器提交手语查询请求时,服务器端会根据各服务器的负载情况将用户手语查询请求分配到合适的服务器上对查询请求进行处理,并且将查询手语图片都的编号以及手语图片所在的云存储节点服务器信息返回给客户端,服务器端的手语图片的云存储方式,有效的实现了负载均衡,不仅提高了用户查询请求的处理效率,同时也使得系统可以利用云技术的特点,具备足够的可扩展性和数据安全性。(2)MapReduce模型改进,实现云节点服务器的负载均衡服务器端在接收到用户请求,并且将图片下载地址返回给客户端的过程中,这个图片下载地址的生成并不是随机的,而是通过对MapReduce计算模型的改进,根据目前云计算服务器节点的工作情况,选择下载任务数量最少的云计算服务器节点中的图片下载URL返回给客户端,从而进一步实现了云计算服务器端的负载均衡,有利于提高整个授予查询系统的性能。(3)采用客户端缓存技术,提高系统性能在客户端将下载的图片,以图片编号为文件名称存储在客户端,在客户向服务器提交查询条件,并且根据服务器所返回的手语图片编号,在客户端文件系统中查询该手语图片,一旦发现在本地文件系统中存在以该图片编号命名的手语图片文件,则不需要再次向云存储服务器节点发送图片下载请求,在用户长时间使用之后,可以较大概率的减少图片下载数量,从而极大的提高了手语查询性能。通过对基于云技术的手语查询系统的模拟实验表明,通过手语图片文件的分布式存储、对MapReduce算法的改进以及客户端的手语图片缓存,在手语图片的存储和方面都具有较好的表现,能快速的响应用户请求。
[Abstract]:With the development of the society and the global integration, the language system for people's communication is more and more abundant, and the sign language system for the hearing-impaired communication is becoming more and more huge and complicated. More and more pictures will be taken to describe and describe sign language, and will be bigger and bigger. For this reason, more and more pictures are described in sign language. How to store and query images in sign language system has become one of the difficulties in sign language development in the information age. In this paper, the cloud storage technology based on Hadoop and mobile terminal technology based on Android are used to provide a sign language graphic query system which can be quickly queried anytime and anywhere, which is of great significance for the application and development of sign language. In this paper, the research of sign language query system based on cloud storage technology mainly includes the following aspects. (1) using cloud technology to improve the performance of sign language query. In this paper, the cloud storage technology is adopted, when the user submits the sign language query request to the server, The server will assign the user sign language query request to the appropriate server to process the query request according to the load of each server. And will query sign language picture number and sign language picture where the cloud storage node server information returned to the client, the server side of the sign language picture cloud storage mode, effectively achieve load balancing, It not only improves the processing efficiency of user query requests, but also enables the system to make use of the characteristics of cloud technology, and has sufficient scalability and data security. (2) the MapReduce model is improved. In the process of receiving the user request and returning the image download address to the client, the load balancing server side of the cloud node server does not generate the image download address randomly. But through the improvement of the MapReduce computing model, according to the current cloud computing server node work situation, select the cloud computing server nodes with the least number of download tasks to download the images of the URL returned to the client. Thus, the load balance of cloud computing server is further realized, which is helpful to improve the performance of the whole grant query system. (3) using the client caching technology to improve the performance of the system to download images in the client, The file name is stored in the client, the query condition is submitted to the server by the client, and the sign language picture is queried in the client file system according to the sign language picture number returned by the server. Once a sign language picture file named after the picture number is found in the local file system, you do not need to send a picture download request to the cloud storage server node again, after the user has used it for a long time, It can reduce the number of image downloads and greatly improve the performance of sign language query. The simulation experiments of sign language query system based on cloud technology show that, through the distributed storage of sign language image files, the improvement of MapReduce algorithm and the client-side sign language picture cache are introduced. In the sign language image storage and good performance, can quickly respond to user requests.
【学位授予单位】：兰州大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP333

【参考文献】