期刊文献+
共找到4篇文章
< 1 >
每页显示 20 50 100
深度学习算法应用于岩石图像处理的可行性研究 被引量:16
1
作者 程国建 刘丽婷 《软件导刊》 2016年第9期163-166,共4页
近年来,机器学习领域中的"深度学习"开始受到国内外学者的广泛关注。借助深度学习,具有多处理层结构的计算机模型可通过多层抽象来学习数据表征。这些方法推动了语音识别、视觉识别、目标检测以及许多其它领域的技术发展。深... 近年来,机器学习领域中的"深度学习"开始受到国内外学者的广泛关注。借助深度学习,具有多处理层结构的计算机模型可通过多层抽象来学习数据表征。这些方法推动了语音识别、视觉识别、目标检测以及许多其它领域的技术发展。深度学习被应用到计算机视觉、语音识别、自然语言处理、在线广告等领域,并取得了良好效果。将计算机技术与地质勘探相结合,在岩石处理方面已取得了不少成果,不断有新的或者改进方法用于提高处理效率和改善效果,以更好地投入到实际生产中去。针对目前传统的岩石薄片图像处理方法,试图将深度学习中典型的两类算法模型:深度信念网络(Deep Belief Networks,DBNs)、卷积神经网络(Convolutional Neural Networks,CNNs)引入到岩石图像的处理中,分析算法在操作中的可行性及优势。 展开更多
关键词 深度学习 深度信念网络 卷积神经网络 岩石图像处理
下载PDF
基于K-Means的颜色量化算法在岩石图像预处理中的应用 被引量:4
2
作者 程国建 魏珺洁 《西安石油大学学报(自然科学版)》 CAS 北大核心 2019年第3期114-119,共6页
提出了一种基于K-Means的颜色量化算法的岩石图像预处理方法,利用K-Means自动将图像中相似的颜色合并为一种颜色,减少了岩石图像中的无用颜色。对于颜色量化,假设RGB岩石图像中的每个像素点都有自己的颜色,并且每个像素点都有对应坐标,... 提出了一种基于K-Means的颜色量化算法的岩石图像预处理方法,利用K-Means自动将图像中相似的颜色合并为一种颜色,减少了岩石图像中的无用颜色。对于颜色量化,假设RGB岩石图像中的每个像素点都有自己的颜色,并且每个像素点都有对应坐标,使用欧几里得距离公式计算每个像素点之间的距离,最后利用K-Means算法对这些有特定颜色的坐标点进行无监督聚类,从而实现颜色的量化。应用结果表明,将基于K-Means的颜色量化算法用在岩石图像中,可以减少岩石图像中的颜色数量,且能在较低性能的计算机设备中很好地再现岩石图像,同时也能提高岩石图像处理的效率。 展开更多
关键词 岩石图像处理 颜色量化 K-MEANS 像素点 聚类
下载PDF
The seismic CT method in measuring rock bodies 被引量:8
3
作者 Du Lizhi Zhang Xiaopei +2 位作者 Niu Jianjun Wang Xuege Feng Guolei 《Applied Geophysics》 SCIE CSCD 2006年第3期192-195,共4页
The seismic computed tomography (CT) method is derived from the basic principles of X-ray section scanning first applied in medical science. The method records P-wave arrivals between shots and receivers in separate... The seismic computed tomography (CT) method is derived from the basic principles of X-ray section scanning first applied in medical science. The method records P-wave arrivals between shots and receivers in separate boreholes. Using the velocity information from 2D seismic P-wave arrival inversion, we can determine the distribution of velocity in rock and soil bodies. This paper introduces a practical case of using the seismic CT method for detecting the structure of the rocks for tunnel engineering and to utilize SIRT algorithms for doing first arrival time iterative inversion. Compared with other exploration methods, it is more efficient and accurate. 展开更多
关键词 seismic computed tomography P-wave velocity travel time inversion.
下载PDF
Quantitative Relationship Between Multi-Angle Polarized Reflectance and BRDF of Rock 被引量:5
4
作者 ZHAO Yun-sheng WU Tai-xia +1 位作者 LIU Bao LUO Yang-jie 《Journal of China University of Mining and Technology》 EI 2005年第3期192-196,共5页
The traditional remote sensing mainly detects the ground vertically to obtain the 2D information but it is hard to get adequate parameters for the quantitative remote sensing to invert land features. The multi-angle o... The traditional remote sensing mainly detects the ground vertically to obtain the 2D information but it is hard to get adequate parameters for the quantitative remote sensing to invert land features. The multi-angle observation can get more detailed and reliable 3D structural parameters of targets, so it makes the quantitative remote sensing applicable. During the process of reflecting, scattering and transmitting the electromagnetic wave, minerals and rocks could reveal the polarized features related to the nature of themselves. Therefore, it has become a new approach of quantitative remote sensing to detect multi-angle polarized information of minerals and rocks. In respect that the polarized reflectance always goes with the bidirectional one, we can obtain the 3D spatial distribution of targets by a polarized means together with detecting its bi-directional reflectance. From the perspective of multi-angle polarized remote sensing mechanism, the quantitative relationship between multi-angle polarized reflectance and the BRDF is studied in this paper. And it is testified that the bi-directional reflectance, polarized reflectance of 45° and the mean value of polarized reflectance are equal to that of the corresponding azimuth angle, zenith angle, detection angle and detection channels in 27t space by experiment. 展开更多
关键词 ROCK Multi-angle Polarized reflectance BRDF Quantitative remote sensing
下载PDF
上一页 1 下一页 到第
使用帮助 返回顶部