摘要
随着数字化技术的不断发展,图像数据量急剧增加,针对从海量的图像数据中检索感兴趣图像效率低的问题,基于Ha-doop云计算平台,采用并行检索技术,利用SIFT尺度不变特征变换算法解决了海量图像检索问题。通过与单机串行检索比较,结果表明基于Hadoop平台的海量图像检索的数据处理能力具有明显优势。
With the continuous development of digital technology,the amount of images data increases drastically.For the problem of inefficient retrieval of interested images from massive images data,in this article,based on Hadoop cloud computing,we resolve the massive image retrieval problem by using parallel retrieval technique and scale-invariant feature transform（SIFT） algorithm.The comparison result with single computer serial retrieval shows that the Hadoop platform-based massive images retrieval has evident superiority in data processing ability.
出处
《计算机应用与软件》
CSCD
北大核心
2013年第5期220-222,共3页
Computer Applications and Software
