The searching method of failure surface which consists of complex geological structures in high and steep rock slopes was studied. Based on computer simulation technology and Monte-Carlo method, three dimensional mult...The searching method of failure surface which consists of complex geological structures in high and steep rock slopes was studied. Based on computer simulation technology and Monte-Carlo method, three dimensional multi-scale geological structures such as engineering scale and statistical scale structures of the slope were simulated. The searching method of failure route which consists of joints and rock bridges was determined via simulation annealing method by considering the shear strength of joints or rock bridges in one supposed route. When shear strengths of all the supposed routes were computed, the least shear strength route was considered failure route. Then, the inclined slice of joint slices and rock bridge slices were separated according to the position of joints and rock bridges. For the rock bridge slices, by distinguishing the failure model, the force direction to the next slice was defined. Finally, the limit equilibrium equations for every slice were established, and the slope stability factor was obtained. One practical example indicates that the discussed method is more closely to the real condition.展开更多
针对岩石薄片图像超分辨率重建过程中因纹理复杂导致现有重建方法效果不理想的问题,提出面向岩石薄片图像的超分辨率网络模型(super-resolution denoising diffusion probability model of rock slice,rsDDPMSR).针对传统上采样方法往...针对岩石薄片图像超分辨率重建过程中因纹理复杂导致现有重建方法效果不理想的问题,提出面向岩石薄片图像的超分辨率网络模型(super-resolution denoising diffusion probability model of rock slice,rsDDPMSR).针对传统上采样方法往往会导致伪影和低分辨率图像先验信息利用不充分的问题提出分层特征增强网络(layered feature enhancement network,LFE-Net),利用双通路网络对平稳小波变换分解后的高频与低频分量进行分层特征增强.为引导扩散模型的生成方向并提供丰富先验信息,将经过LFE-Net增强后的低分辨率特征与目标高分辨率加噪图像特征通道拼接作为扩散模型的条件输入.在U-Net的基础上设计了双编码器多尺度噪声预测网络(ACA-U-Net)有效处理岩石薄片多尺度信息并在跳跃连接中引入时间感知的自适应交叉注意力机制适配扩散模型不同去噪阶段的特征分布变化增强模型对关键区域的关注程度,有效提升图像重建细节.实验结果表明,rsDDPMSR在2×、4×、8×放大倍数下,峰值信噪比(PSNR)和结构相似度(SSIM)相比于CAMixerSR、SDFlow、IDM和SR3等主流重建方法具有更优的重建效果.展开更多
The</span><span style="font-family:""> </span><span style="font-family:Verdana;">western part of north Tarim Uplift underwent multi-stage tectonic movement and multiple...The</span><span style="font-family:""> </span><span style="font-family:Verdana;">western part of north Tarim Uplift underwent multi-stage tectonic movement and multiple stages of magmatism.</span><span style="font-family:""> </span><span style="font-family:Verdana;">Igneous rocks are associated with carbonate and buried deep.</span><span style="font-family:""> </span><span style="font-family:Verdana;">The seismic response characteristics of igneous rocks are similar in many respects to the seismic response characteristics of karst, making the identification and prediction of igneous rocks more difficult.</span><span style="font-family:""> </span><span style="font-family:Verdana;">This study compares the seismic reflection characteristics of igneous rocks. Set up three types of igneous rock seismic facies model penetration type, fracture type and central type</span><span style="font-family:Verdana;">. </span><span style="font-family:Verdana;">And it concluded that a time-slice, coherence analysis, analysis of the seismic properties of the layers and the method of three-dimensional engraving to identify the igneous rocks. This method has been applied to the identification and prediction of carbonate rock buried </span><span style="font-family:Verdana;">and </span><span style="font-family:Verdana;">hill igneous rocks in the north Tarim basin YingMaiLi region and has achieved good results.展开更多
岩石薄片图像对研究石油地质特性以及油气勘探都有重要的意义。由于各种因素的限制,获取到的岩石薄片图像经常会出现分辨率较低的情况,一定程度上限制了研究者对其细节信息的掌握。而一般的神经网络超分辨率算法都需要大量的数据作为训...岩石薄片图像对研究石油地质特性以及油气勘探都有重要的意义。由于各种因素的限制,获取到的岩石薄片图像经常会出现分辨率较低的情况,一定程度上限制了研究者对其细节信息的掌握。而一般的神经网络超分辨率算法都需要大量的数据作为训练集,为了提升岩石薄片图像超分辨率重建算法纹理细节信息还原能力,本文利用单图像生成式对抗网络,不需输入大量数据集,对岩石薄片图像进行超分辨率重建。采用鄂尔多斯某油区岩石铸体薄片图像进行训练,通过峰值信噪比(Peak Signal to Noise Ratio,SSNR)和结构相似性(Structural Similarity,SSIM)评价指标进行模型评价,实验结果表明:该方法超分辨率处理的图像在视觉效果和评价指标上均具有良好的效果。展开更多
基金Project(50539100) supported by the National Natural Science Foundation of ChinaProject(BK2006171) supported by the Jiangsu Natural Science Foundation
文摘The searching method of failure surface which consists of complex geological structures in high and steep rock slopes was studied. Based on computer simulation technology and Monte-Carlo method, three dimensional multi-scale geological structures such as engineering scale and statistical scale structures of the slope were simulated. The searching method of failure route which consists of joints and rock bridges was determined via simulation annealing method by considering the shear strength of joints or rock bridges in one supposed route. When shear strengths of all the supposed routes were computed, the least shear strength route was considered failure route. Then, the inclined slice of joint slices and rock bridge slices were separated according to the position of joints and rock bridges. For the rock bridge slices, by distinguishing the failure model, the force direction to the next slice was defined. Finally, the limit equilibrium equations for every slice were established, and the slope stability factor was obtained. One practical example indicates that the discussed method is more closely to the real condition.
文摘针对岩石薄片图像超分辨率重建过程中因纹理复杂导致现有重建方法效果不理想的问题,提出面向岩石薄片图像的超分辨率网络模型(super-resolution denoising diffusion probability model of rock slice,rsDDPMSR).针对传统上采样方法往往会导致伪影和低分辨率图像先验信息利用不充分的问题提出分层特征增强网络(layered feature enhancement network,LFE-Net),利用双通路网络对平稳小波变换分解后的高频与低频分量进行分层特征增强.为引导扩散模型的生成方向并提供丰富先验信息,将经过LFE-Net增强后的低分辨率特征与目标高分辨率加噪图像特征通道拼接作为扩散模型的条件输入.在U-Net的基础上设计了双编码器多尺度噪声预测网络(ACA-U-Net)有效处理岩石薄片多尺度信息并在跳跃连接中引入时间感知的自适应交叉注意力机制适配扩散模型不同去噪阶段的特征分布变化增强模型对关键区域的关注程度,有效提升图像重建细节.实验结果表明,rsDDPMSR在2×、4×、8×放大倍数下,峰值信噪比(PSNR)和结构相似度(SSIM)相比于CAMixerSR、SDFlow、IDM和SR3等主流重建方法具有更优的重建效果.
文摘The</span><span style="font-family:""> </span><span style="font-family:Verdana;">western part of north Tarim Uplift underwent multi-stage tectonic movement and multiple stages of magmatism.</span><span style="font-family:""> </span><span style="font-family:Verdana;">Igneous rocks are associated with carbonate and buried deep.</span><span style="font-family:""> </span><span style="font-family:Verdana;">The seismic response characteristics of igneous rocks are similar in many respects to the seismic response characteristics of karst, making the identification and prediction of igneous rocks more difficult.</span><span style="font-family:""> </span><span style="font-family:Verdana;">This study compares the seismic reflection characteristics of igneous rocks. Set up three types of igneous rock seismic facies model penetration type, fracture type and central type</span><span style="font-family:Verdana;">. </span><span style="font-family:Verdana;">And it concluded that a time-slice, coherence analysis, analysis of the seismic properties of the layers and the method of three-dimensional engraving to identify the igneous rocks. This method has been applied to the identification and prediction of carbonate rock buried </span><span style="font-family:Verdana;">and </span><span style="font-family:Verdana;">hill igneous rocks in the north Tarim basin YingMaiLi region and has achieved good results.
文摘岩石薄片图像对研究石油地质特性以及油气勘探都有重要的意义。由于各种因素的限制,获取到的岩石薄片图像经常会出现分辨率较低的情况,一定程度上限制了研究者对其细节信息的掌握。而一般的神经网络超分辨率算法都需要大量的数据作为训练集,为了提升岩石薄片图像超分辨率重建算法纹理细节信息还原能力,本文利用单图像生成式对抗网络,不需输入大量数据集,对岩石薄片图像进行超分辨率重建。采用鄂尔多斯某油区岩石铸体薄片图像进行训练,通过峰值信噪比(Peak Signal to Noise Ratio,SSNR)和结构相似性(Structural Similarity,SSIM)评价指标进行模型评价,实验结果表明:该方法超分辨率处理的图像在视觉效果和评价指标上均具有良好的效果。
