We introduce a new approach to image super-resolution. The idea is to use a simple wavelet-based linear interpolation scheme as our initial estimate of high-resolution image;and to intensify geometric structure in ini...We introduce a new approach to image super-resolution. The idea is to use a simple wavelet-based linear interpolation scheme as our initial estimate of high-resolution image;and to intensify geometric structure in initial estimation with an iterative projection process based on hard-thresholding scheme in a new angular multiselectivity domain. This new domain is defined by combining of laplacian pyramid and angular multiselectivity decomposition, the result is multiselective contourlets which can capture and restore adaptively and slightly better geometric structure of image. The experimental results demonstrate the effectiveness of the proposed approach.展开更多
在遮挡目标识别中,目标可能会被其他物体遮挡,导致目标的部分有效特征丢失或变形。目标有效特征的减少,使得单一YOLOv4(You Only Look Once version 4)无法准确识别锚框的初始值,使得模型目标识别困难。为此,引入K-means++算法改进单一Y...在遮挡目标识别中,目标可能会被其他物体遮挡,导致目标的部分有效特征丢失或变形。目标有效特征的减少,使得单一YOLOv4(You Only Look Once version 4)无法准确识别锚框的初始值,使得模型目标识别困难。为此,引入K-means++算法改进单一YOLOv4算法,提出基于改进YOLOv4的遮挡目标识别算法。通过非下采样Contourlet变换划分图像为低频部分和高频部分,分别利用线性增强函数和改进的自适应阈值增强图像,并经由非下采样Contourlet逆变换生成重建图像,对其执行模糊对比度增强。选取YOLOv4作为目标识别基础模型,采用深度可分离卷积替代模型中部分卷积,并替换特征金字塔为递归特征金字塔,提升小目标和遮挡目标的特征学习能力。引入K-means++算法自适应获取锚框,优化锚框初始值,并利用完全交并比和交叉熵构建损失函数,训练改进的YOLOv4并将增强后图像输入其中,实现遮挡目标识别。实验结果表明,所提方法能够有效识别图像目标,且识别结果P-R曲线更理想。展开更多
遥感影像在采集过程中,地面覆盖种类数量庞大且采集影像清晰度低、分辨率较差,关键像素特征之间的阈值衡量标准模糊,导致信息提取难度增大,从而降低信息利用率。为此,提出了基于像素紧密程度的多源遥感影像信息提取方法。利用Contourle...遥感影像在采集过程中,地面覆盖种类数量庞大且采集影像清晰度低、分辨率较差,关键像素特征之间的阈值衡量标准模糊,导致信息提取难度增大,从而降低信息利用率。为此,提出了基于像素紧密程度的多源遥感影像信息提取方法。利用Contourlet变换,实现遥感影像空间域、变换域的多角度增强,优化遥感影像整体清晰度。利用简单线性迭代聚类(Simple Linear Iterative Clustering,SLIC)超像素算法计算像素聚类中心与邻近像素紧密程度,摆脱固定阈值影响。引入灰度共生矩阵(Gray-level Co-occurrenceMatrix,GLCM),提取主体特征信息;构建相关向量机分类模型,结合拉普拉斯二次逼近回归算法将提取问题转化为噪声回归问题,并对其展开求解,进而实现遥感影像的信息提取。实验结果表明:所提方法对遥感信息主体的分类与真实遥感信息主体分类基本一致,在信息提取过程中的错提取率和漏提取率低,总体提取精度保持在99%以上,且对道路信息提取清晰度高,表明该方法提高了遥感信息的解译水平。展开更多
Currently, horizontal well fracturing is indispensable for shale gas development. Due to the variable reservoir formation morphology, the drilling trajectory often deviates from the high-quality reservoir,which increa...Currently, horizontal well fracturing is indispensable for shale gas development. Due to the variable reservoir formation morphology, the drilling trajectory often deviates from the high-quality reservoir,which increases the risk of fracturing. Accurately recognizing low-amplitude structures plays a crucial role in guiding horizontal wells. However, existing methods have low recognition accuracy, and are difficult to meet actual production demand. In order to improve the drilling encounter rate of high-quality reservoirs, we propose a method for fine recognition of low-amplitude structures based on the non-subsampled contourlet transform(NSCT). Firstly, the seismic structural data are analyzed at multiple scales and directions using the NSCT and decomposed into low-frequency and high-frequency structural components. Then, the signal of each component is reconstructed to eliminate the low-frequency background of the structure, highlight the structure and texture information, and recognize the low-amplitude structure from it. Finally, we combined the drilled horizontal wells to verify the low-amplitude structural recognition results. Taking a study area in the west Sichuan Basin block as an example, we demonstrate the fine identification of low-amplitude structures based on NSCT. By combining the variation characteristics of logging curves, such as organic carbon content(TOC), natural gamma value(GR), etc., the real structure type is verified and determined, and the false structures in the recognition results are checked. The proposed method can provide reliable information on low-amplitude structures for optimizing the trajectory of horizontal wells. Compared with identification methods based on traditional wavelet transform and curvelet transform, NSCT enhances the local features of low-amplitude structures and achieves finer mapping of low-amplitude structures, showing promise for application.展开更多
文摘We introduce a new approach to image super-resolution. The idea is to use a simple wavelet-based linear interpolation scheme as our initial estimate of high-resolution image;and to intensify geometric structure in initial estimation with an iterative projection process based on hard-thresholding scheme in a new angular multiselectivity domain. This new domain is defined by combining of laplacian pyramid and angular multiselectivity decomposition, the result is multiselective contourlets which can capture and restore adaptively and slightly better geometric structure of image. The experimental results demonstrate the effectiveness of the proposed approach.
文摘在遮挡目标识别中,目标可能会被其他物体遮挡,导致目标的部分有效特征丢失或变形。目标有效特征的减少,使得单一YOLOv4(You Only Look Once version 4)无法准确识别锚框的初始值,使得模型目标识别困难。为此,引入K-means++算法改进单一YOLOv4算法,提出基于改进YOLOv4的遮挡目标识别算法。通过非下采样Contourlet变换划分图像为低频部分和高频部分,分别利用线性增强函数和改进的自适应阈值增强图像,并经由非下采样Contourlet逆变换生成重建图像,对其执行模糊对比度增强。选取YOLOv4作为目标识别基础模型,采用深度可分离卷积替代模型中部分卷积,并替换特征金字塔为递归特征金字塔,提升小目标和遮挡目标的特征学习能力。引入K-means++算法自适应获取锚框,优化锚框初始值,并利用完全交并比和交叉熵构建损失函数,训练改进的YOLOv4并将增强后图像输入其中,实现遮挡目标识别。实验结果表明,所提方法能够有效识别图像目标,且识别结果P-R曲线更理想。
文摘遥感影像在采集过程中,地面覆盖种类数量庞大且采集影像清晰度低、分辨率较差,关键像素特征之间的阈值衡量标准模糊,导致信息提取难度增大,从而降低信息利用率。为此,提出了基于像素紧密程度的多源遥感影像信息提取方法。利用Contourlet变换,实现遥感影像空间域、变换域的多角度增强,优化遥感影像整体清晰度。利用简单线性迭代聚类(Simple Linear Iterative Clustering,SLIC)超像素算法计算像素聚类中心与邻近像素紧密程度,摆脱固定阈值影响。引入灰度共生矩阵(Gray-level Co-occurrenceMatrix,GLCM),提取主体特征信息;构建相关向量机分类模型,结合拉普拉斯二次逼近回归算法将提取问题转化为噪声回归问题,并对其展开求解,进而实现遥感影像的信息提取。实验结果表明:所提方法对遥感信息主体的分类与真实遥感信息主体分类基本一致,在信息提取过程中的错提取率和漏提取率低,总体提取精度保持在99%以上,且对道路信息提取清晰度高,表明该方法提高了遥感信息的解译水平。
基金supported by Sichuan Science and Technology Program under Grant 2024NSFSC1984 and Grant 2024NSFSC1990。
文摘Currently, horizontal well fracturing is indispensable for shale gas development. Due to the variable reservoir formation morphology, the drilling trajectory often deviates from the high-quality reservoir,which increases the risk of fracturing. Accurately recognizing low-amplitude structures plays a crucial role in guiding horizontal wells. However, existing methods have low recognition accuracy, and are difficult to meet actual production demand. In order to improve the drilling encounter rate of high-quality reservoirs, we propose a method for fine recognition of low-amplitude structures based on the non-subsampled contourlet transform(NSCT). Firstly, the seismic structural data are analyzed at multiple scales and directions using the NSCT and decomposed into low-frequency and high-frequency structural components. Then, the signal of each component is reconstructed to eliminate the low-frequency background of the structure, highlight the structure and texture information, and recognize the low-amplitude structure from it. Finally, we combined the drilled horizontal wells to verify the low-amplitude structural recognition results. Taking a study area in the west Sichuan Basin block as an example, we demonstrate the fine identification of low-amplitude structures based on NSCT. By combining the variation characteristics of logging curves, such as organic carbon content(TOC), natural gamma value(GR), etc., the real structure type is verified and determined, and the false structures in the recognition results are checked. The proposed method can provide reliable information on low-amplitude structures for optimizing the trajectory of horizontal wells. Compared with identification methods based on traditional wavelet transform and curvelet transform, NSCT enhances the local features of low-amplitude structures and achieves finer mapping of low-amplitude structures, showing promise for application.
