检索结果-南通市图书馆

Optimization of the seismic processing phase-shift plus finite-difference migration operator based on a hybrid genetic and simulated annealing algorithm

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Petroleum Science 2013年第2期10卷 190-194页

作者： luo renze Huang Yuanyi Liang Xianghao luo Jun Cao Ying State Key Laboratory of Oil and Gas Reservoir Geology and Exploitation Southwest Petroleum University Geophysics Development Department of BGP Exploration and Exploitation Department of Tarim Oilfield Company PetroChina Changqing Oilfield Company PetroChina

Although the phase-shift seismic processing method has characteristics of high accuracy, good stability, high efficiency, and high-dip imaging, it is not able to adapt to strong lateral velocity variation. To overcome this defect, a finite-difference method in the frequency-space domain is introduced in the migration process, because it can adapt to strong lateral velocity variation and the coefficient is optimized by a hybrid genetic and simulated annealing algorithm. The two measures improve the precision of the approximation dispersion equation. Thus, the imaging effect is improved for areas of high-dip structure and strong lateral velocity variation. The migration imaging of a 2-D SEG/EAGE salt dome model proves that a better imaging effect in these areas is achieved by optimized phase-shift migration operator plus a finite-difference method based on a hybrid genetic and simulated annealing algorithm. The method proposed in this paper is better than conventional methods in imaging of areas of high-dip angle and strong lateral velocity variation.

关键词： Migration operator phase-shift plus finite-difference hybrid algorithm genetic andsimulated annealing algorithm optimization coefficient

REDUCING THE PEAK-TO-AVERAGE POWER RATIO OF OFDM SYSTEM WITH LOW COMPLEXITY

在线全文

学校读者我要写书评

暂无评论

Journal of Electronics(China) 2006年第1期23卷 26-28页

作者： luo renze Jing Longjiang luo Lang Li Jie Zhu Weile Dept of Electro. Eng. Univ. of Electronic Science and Technology of China Chengdu 610054 China Zhongshan Institute Univ. of Electronic Science and Technology of China Zhongshan 528402 China

This paper proposes a companding scheme, where small signals are enlarged and large signals are reduced, to reduce the Peak-to-Average Power Ratio(PAPR). Computer simulation results show that the proposed technique has two advantages at least when compared with the conventional methods such as partial transmit sequence, selective mapping and the previous companding. First, it gets better PAPR performances with a lower complexity. Second, the scheme achieves greater performances gain with hardly any damnification of OFDM signals in some degree.

关键词：正交频分复用 OFDM 功率比率 PAPR 压缩扩展器

在线全文

学校读者我要写书评

暂无评论

基于ACGAN和迁移学习的骨显像分类方法

光学精密工程 2023年第6期31卷 936-949页

作者：余泓罗仁泽陈春梦唐祥罗任权西南石油大学电气信息学院四川成都610500 宜宾市第二人民医院核医学科四川宜宾644000 西南石油大学计算机科学学院四川成都610500

由于骨显像存在样本数量有限、类别不平衡的问题,导致骨显像分类存在较大困难。为提升骨显像的分类准确率,本文提出了一种基于结合辅助分类器的生成对抗网络(ACGAN)数据生成和迁移学习的骨显像分类方法。首先,为解决骨显像类别不平衡的... 详细信息

由于骨显像存在样本数量有限、类别不平衡的问题,导致骨显像分类存在较大困难。为提升骨显像的分类准确率,本文提出了一种基于结合辅助分类器的生成对抗网络(ACGAN)数据生成和迁移学习的骨显像分类方法。首先,为解决骨显像类别不平衡的问题,设计了一种MU-ACGAN模型。该模型以U-Net为生成器框架,同时结合密集残差连接和通道-空间注意力机制结构来提升骨显像细节特征生成,判别器通过密集残差注意力卷积块提取骨显像特征进行判别;然后,结合传统数据增强方式进一步扩充数据量;最后,设计了一种多尺度卷积神经网络提取骨显像不同尺度的特征,提升分类效果。在模型训练过程中,采用两阶段迁移学习方式,优化模型的初始化参数、解决过拟合的问题。实验结果表明,本文提出方法分类准确率达到了85.71%,有效缓解了小样本骨显像数据集分类准确率不高的问题。

关键词：骨显像结合辅助分类器的生成对抗网络(ACGAN) 迁移学习注意力机制数据增强

基于U-Net的两阶段SPECT骨显像降噪方法研究

在线全文

学校读者我要写书评

暂无评论

光电子．激光 2023年第7期34卷 771-784页

作者：余泓罗仁泽陈春梦郭亮罗任权西南石油大学电气信息学院四川成都610500 宜宾市第二人民医院核医学科四川宜宾644000 西南石油大学地球科学与技术学院四川成都610500

进行单光子发射计算机断层成像(single-photon emission computed tomography,SPECT)骨显像检查时,为减少给病人带来的辐射伤害,医师常会减轻辐射剂量,导致骨显像信噪比、分辨率较低,严重影响诊断以及病灶自动检测效果。为提升骨显像质... 详细信息

进行单光子发射计算机断层成像(single-photon emission computed tomography,SPECT)骨显像检查时,为减少给病人带来的辐射伤害,医师常会减轻辐射剂量,导致骨显像信噪比、分辨率较低,严重影响诊断以及病灶自动检测效果。为提升骨显像质量,提出了一种基于U-Net的两阶段SPECT骨显像降噪方法。首先,设计了一种U-Net噪声估计网络来快速估计每张骨显像的噪声水平,为主干降噪网络提供噪声先验知识。其次,主干降噪网络同样以U-Net为基础框架,同时结合多尺度特征融合、通道-空间注意力机制结构来增强网络的噪声特征提取能力,预测出噪声图。最后,通过残差学习得到降噪骨显像。同时,为解决使用均方误差(mean square error,MSE)损失函数的重建图像过于平滑的问题,设计了一种复合损失函数,保留骨显像的原有细节信息。实验中,向训练集中的骨显像施加不同噪声水平进行数据扩充,并且采用迁移策略解决模型过拟合问题。结果表明,与目前主流算法相比,所提出的降噪方法能够有效降低骨显像噪声,并且保留病灶细节特征。此外,通过盲降噪能够改善原骨显像质量、提升病灶自动分割效果。

关键词： SPECT骨显像深度学习噪声水平注意力机制残差学习

基于改进U-Net的SPECT骨显像病灶分割研究

在线全文

学校读者我要写书评

暂无评论

光电子．激光 2022年第10期33卷 1110-1120页

作者：余泓罗仁泽陈春梦罗任权李华督西南石油大学电气信息学院四川成都610500 宜宾市第二人民医院核医学科四川宜宾644000

在核医学中,单光子发射计算机断层(single-photon emission computed tomograpy,SPECT)骨显像是辅助医师诊断癌症的重要手段。针对骨显像图像信噪比低、边界模糊、病灶小难以提取和人工勾画病灶耗时等问题,提出一种基于改进U-Net网络的... 详细信息

在核医学中,单光子发射计算机断层(single-photon emission computed tomograpy,SPECT)骨显像是辅助医师诊断癌症的重要手段。针对骨显像图像信噪比低、边界模糊、病灶小难以提取和人工勾画病灶耗时等问题,提出一种基于改进U-Net网络的骨显像病灶自动分割算法。该算法在U-Net的原卷积块基础上,采用了多尺度密集连接(multi-scale dense connection,MDC)的方式来提高对小病灶特征的提取能力,同时解决了网络加深后出现的梯度消失问题。其次,为提取病灶的细节特征,在密集连接和跳跃连接处引入了注意力机制结构。最后,针对使用小样本数据集,模型难以收敛的问题,采用迁移学习的方法,优化了模型的初始参数,提升模型的泛化能力和分割效率。此外,为了降低计算量、进一步提高分割效果,对数据集进行了裁剪和去噪。同时,将处理后的图像采用旋转、镜像等方法进行了数据扩充。实验结果表明,改进的U-Net的识别精确率(precision)、平均交并比(mean intersection-over-union,mIoU)分别能达到0.7352、0.4673,效果优于目前主流的分割算法,具有一定实际应用价值。

关键词： SPECT骨显像多尺度密集连接(MDC) 图像处理注意力机制迁移学习

在线全文

学校读者我要写书评

暂无评论

基于FE-Unet的机场道面裂缝检测

光电子．激光 2023年第1期34卷 34-42页

作者：邓治林罗仁泽费越李海丰西南石油大学电气信息学院四川成都610500 成都圭目机器人有限公司四川成都610101 中国民航大学计算机科学与技术学院天津300300

机场道面裂缝具有形态复杂多变、走向不连续、数据噪音多等特征,现有算法模型均未达到令人满意的结果。为了改善裂缝检测效果,本文提出了一种新的深度学习模型,命名为“FE-Unet”。该模型采用改进的残差连接方式,在解决多层网络下梯度... 详细信息

机场道面裂缝具有形态复杂多变、走向不连续、数据噪音多等特征,现有算法模型均未达到令人满意的结果。为了改善裂缝检测效果,本文提出了一种新的深度学习模型,命名为“FE-Unet”。该模型采用改进的残差连接方式,在解决多层网络下梯度的回传问题的同时起到细化特征以及整合通道的信息作用,提升了各阶段特征的区分度;此外,模型中的通道注意力模块(channel attention block, CAB)可以更好地提取判别特征,增强预测的一致性;最后,利用焦点损失(focal loss, FL)使模型专注于难分类的细小裂缝。实验中,以实际7 778张机场道面裂缝图像来训练模型,并在1 701张图像上进行验证。在与经典的全卷积神经网络(fully convolutional network, FCN)、DeepLab v3和Unet对比实验中,FE-Unet对裂缝、灌缝和板缝的检测性能均优于其他模型。其中,FE-Unet对裂缝检测的精度、召回率、F1值分别达到了80.31%、82.72%和81.49%。

关键词：深度学习图像处理机场道面检测 FE-Unet 裂缝识别

OFDM Wireless Downhole Transmission Systems and Proposed SLM Method for PAPR Reduction

在线全文

学校读者我要写书评

暂无评论

Communications and Network 2014年第1期6卷 9-14页

作者： renze luo Yupu Dang Rui Li Jiao Yang Department of Well Tech China Oilfield Services Limited Beijing China State Key Laboratory of Oil and Gas Reservoir Geology and Exploitation Southwest Petroleum University Chengdu China

At present, mud pulse transmission systems are widely used in downhole data transmission. But the systems are very low in transmission efficiency, only 5-10 bits/s, with very large anti-inter-symbol-interference (ISI). It cannot meet high requirements for high-speed transmission of modern logging system. The development of communication technology has laid some foundation for this requirement. For this purpose, the Orthogonal Frequency Division Multiplexing (OFDM) Wireless Downhole Transmission Systems are proposed for the first time because of their high transmission rate, anti-inter-symbol-interference (ISI), and high spectral efficiency, etc. Due to non-linear power amplifier (PA) of logging systems with limited dynamic range, the drawbacks of high peak-average power ratio (PAPR) may outweigh all the potential benefits of OFDM wireless downhole transmission systems. Selective mapping (SLM) method can reduce the PAPR of OFDM logging signals without distortion. But at the receiver, the conventional SLM method needs exact bits of side information (SI) to recover the data signal. The probability of erroneous SI detection has a significant influence on the error performance of the system. And individual transmissions of SI result in the reduction of bandwidth efficiency. To restore the exact data signal, our scheme codes the SI bits by linear block codes (LBC), and is easily decoded by syndrome decoding. And then the coding SI bits are superimposed onto the logging signals to omit SI bits transmission. The theory and simulation results show that the proposed method has better performance than the conventional one. Accordingly, the OFDM wireless downhole transmission systems can tackle the high PAPR problem, and highten the transmission rate of logging signals.

关键词： Orthogonal Frequency Division Multiplexing Wireless Downhole Transmission Systems Orthogonal Frequency Division Multiplexing (OFDM) Logging While Drilling (LWD) Peak-to-Average Power Ratio (PAPR) Selected Mapping (SLM)

A PTS Optimization Scheme with Superimposed Training for PAPR Reduction in OFDM System

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Communications and Network 2014年第2期6卷 97-104页

作者： renze luo Rui Li Xiaoqiong Wu Shuainan Hu Na Niu State Key Laboratory of Oil and Gas Reservoir Geology and Exploitation Southwest Petroleum University Chengdu Sichuan China Zhongshan Branch China Telecom Co. Ltd. Zhongshan Guangdong China

Partial Transmit Sequences (PTS) is an efficient scheme for Peak-to-Average Power Ratio (PAPR) reduction in Orthogonal Frequency Division Multiplexing (OFDM) system. It does not bring any signal distortion. However, its remarkable drawback is the high computational complexity. In order to reduce the computational complexity, currently many PTS methods have been proposed but with the cost of the loss of PAPR performance of the system. In this paper, we introduce an improved PTS optimization method with superimposed training. Simulation results show that, compared with conventional PTS, improved PTS scheme can achieve better PAPR performance while be implemented with lower computation complexity of the system.

关键词： Orthogonal Frequency Division Multiplexing (OFDM) Peak-to-Average Power Ratio (PAPR) Partial Transmit Sequences (PTS) Superimposed Training

基于改进Faster RCNN的射线图像焊缝缺陷检测方法

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

电子测量技术 2023年第22期46卷 160-168页

作者：罗仁泽唐祥余泓李华督西南石油大学计算机科学学院成都610500 西南石油大学电气信息学院成都610500

针对X射线图像中小目标缺陷检测和多尺寸缺陷检测的问题,提出一种基于改进Faster RCNN的焊缝缺陷检测算法。首先,该算法采用ResNet50、特征金字塔网络作为Faster RCNN检测网络的主干网络,达到在多个特征图上检测不同尺寸缺陷的目的;然... 详细信息

针对X射线图像中小目标缺陷检测和多尺寸缺陷检测的问题,提出一种基于改进Faster RCNN的焊缝缺陷检测算法。首先,该算法采用ResNet50、特征金字塔网络作为Faster RCNN检测网络的主干网络,达到在多个特征图上检测不同尺寸缺陷的目的;然后在主干网络前增加背景减去网络层,来降低图像背景对小目标缺陷的干扰;接着利用三支路区域推荐网络层细化原始区域推荐网络层中候选框的预测任务,从而减少候选框数量、优化检测速度;最后对网络中卷积层的数量进行微调,增强网络特征提取能力。实验结果表明,改进网络的均值平均精度和每张图像检测速度分别为83.09%、20.8 ms,相比改进前的网络,预设的锚框增加了10779个,检测速度仅仅降低了3.1 ms,均值平均精度提高了19.43%。改进网络在保证检测速度的基础上,有效提高了对小目标缺陷和多尺寸缺陷的检测效果。

关键词：焊缝缺陷深度学习目标检测 Faster RCNN