卷积神经网络在岩石薄片图像检索中的应用初探

Feasibility study of Convolutional Neural Network in image retrieval of rock slices

近几年来,卷积神经网络引起了国内外研究者的广泛关注,并在大规模图像处理方面有出色的表现,尤其在模式识别领域。将地质勘探与计算机技术相结合,在岩石图像处理方面已经取得了较好的成绩,并且还在不断的探索中,以求更好地投入到实际中去。对于地质勘探研究者来说,对于大量的岩石薄片图像,如何进行快速并且有效的检索是值得研究的领域课题。传统的基于文本的检索方式已不能满足要求,为此,本文试图将卷积神经网络引入到岩石薄片图像的检索中,分析其在岩石薄片图像检索中的可行性。

In recent years,Convolutional Neural Network has attracted wide attention of researchers at home and abroad,and it has excellent performance in large-scale image processing,especially in the field of pattern recognition.The combination of geological exploration and computer technology has achieved good results in the process of rock image processing,and it is still in constant exploration in order to better put into practice. For geological exploration researchers,it is a necessary research area for a large number of images of rock flakes to be retrieved quickly and efficiently. Traditional text-based retrieval methods can ＇t meet the requirements,this paper attempts to introduce Convolution Neural Network into the rock slice image retrieval,and analyze its feasibility in the rock slice image retrieval.