大数据分析技术在海量激光图像检索中的应用

Application of big data analysis technology in massive laser image retrieval

为解决海量激光图像的检索问题,提出基于大数据分析的海量激光图像检索方法。大数据分析技术中的HDFS采用一个NameNode与数个DataNode组合的运行方式,将海量激光图像特征数据存储在HDFS中;采用基于MapReduce程序与相似度计算方法检索海量激光图像,利用Map读取存储在HDFS中的海量激光图像特征数据,并转换为Key/Value形式的数据特征,在Reduce任务中采用相似度计算方法获取HDFS中海量激光图像特征的形状、纹理、颜色相似度并排序,排序结果则为激光图像检索结果。实验结果表明,所提方法计算的激光图像特征相似度误差较小,检索海量激光图像过程中,所提方法具有高查全率、高查准率的优势,是一种高性能的海量激光图像检索方法。

In order to solve the retrieval problem of massive laser image,a massive laser image retrieval method based on large data analysis is proposed.HDFS in big data analysis technology uses a combination function method of NameNode and several DataNode to store massive laser image feature data in HDFS.The MapReduce program and similarity calculation methods are adopted to retrieve massive laser images.The Map is used to read massive laser image feature data stored in HDFS,then convert to Key/Value data features.In the Reduse mission.Similarity calculation method is used to obtain the shape,texture,and color similarity of the massive laser image features in HDFS,and the sorting results are the laser image retrieval results.The experimental results show that the proposed method has the advantages of Gaocha’s full rate and Gaocha’s accurate rate,and is a high performance laser image retrieval method.