基于Hadoop分布式的海量图像检索方法研究

Research on Mass Image Retrieval Method Based on Hadoop Distributed

为解决传统海量图像检索方法存在的问题,本文提出基于Hadoop分布式的海量图像检索方法.该方法利用Hadoop的聚类功能分布式存储海量图像,确定海量图像特征的模糊空间,计算海量图像特征相似度,以海量图像特征检索匹配相似度为核心指标,得出基于Hadoop分布式检索海量图像结果,完成海量图像检索。实例结果表明:设计检索方法测得的检索查全率高于对照组,能够解决检索查全率低的问题。

In order to solve the problems of traditional mass image retrieval methods,this paper proposes a distributed mass image retrieval method based on Hadoop.This method uses the clustering function of Hadoop to store massive images in a distributed manner,determines the fuzzy space of massive image features,calculates the similarity of massive image features,and uses the similarity of massive image feature retrieval and matching as the core index to obtain distributed retrieval of massive images based on Hadoop As a result,massive image retrieval is completed.The example results show that the retrieval recall rate measured by the designed retrieval method is higher than that of the control group,which can solve the problem of low retrieval recall rate.