The transient flow mathematical model of arbitrary shaped heterogeneous reservoirs with impermeability barrier is proposed in this paper. In order to establish this model, the perturbation method is employed and the s...The transient flow mathematical model of arbitrary shaped heterogeneous reservoirs with impermeability barrier is proposed in this paper. In order to establish this model, the perturbation method is employed and the solution of model is expanded into a series in powers of perturbation parameter. By using the Boundary Element Method (BEM) and Duhamel principle, wellbore pressure with effects of skins and wellbore storage is obtained. The type curves are plotted and analyzed considering effects of heterogeneity, arbitrary shape and impermeable barriers. Finally, the results obtained by perturbation boundary element method is compared with the analytical solution and is available for the transient pressure analysis of arbitrary shaped reservoirs.展开更多
目的获取场景图像中的文本信息对理解场景内容具有重要意义,而文本检测是文本识别、理解的基础。为了解决场景文本识别中文字定位不准确的问题,本文提出了一种高效的任意形状文本检测器:非局部像素聚合网络。方法该方法使用特征金字塔...目的获取场景图像中的文本信息对理解场景内容具有重要意义,而文本检测是文本识别、理解的基础。为了解决场景文本识别中文字定位不准确的问题,本文提出了一种高效的任意形状文本检测器:非局部像素聚合网络。方法该方法使用特征金字塔增强模块和特征融合模块进行轻量级特征提取,保证了速度优势;同时引入非局部操作以增强骨干网络的特征提取能力,使其检测准确性得以提高。非局部操作是一种注意力机制,能捕捉到文本像素之间的内在关系。此外,本文设计了一种特征向量融合模块,用于融合不同尺度的特征图,使尺度多变的场景文本实例的特征表达得到增强。结果本文方法在3个场景文本数据集上与其他方法进行了比较,在速度和准确度上均表现突出。在ICDAR(International Conference on Document Analysis and Recognition)2015数据集上,本文方法比最优方法的F值提高了0.9%,检测速度达到了23.1帧/s;在CTW(Curve Text in the Wild)1500数据集上,本文方法比最优方法的F值提高了1.2%,检测速度达到了71.8帧/s;在Total-Text数据集上,本文方法比最优方法的F值提高了1.3%,检测速度达到了34.3帧/s,远远超出其他方法。结论本文方法兼顾了准确性和实时性,在准确度和速度上均达到较高水平。展开更多
基金Project supported by the National Natural Science Foundation of China (Grant Nos: 50174011 10172028) the Science Tech. Research Program of Heilongjiang Provincial Education Department (Grant No: 10531032).
文摘The transient flow mathematical model of arbitrary shaped heterogeneous reservoirs with impermeability barrier is proposed in this paper. In order to establish this model, the perturbation method is employed and the solution of model is expanded into a series in powers of perturbation parameter. By using the Boundary Element Method (BEM) and Duhamel principle, wellbore pressure with effects of skins and wellbore storage is obtained. The type curves are plotted and analyzed considering effects of heterogeneity, arbitrary shape and impermeable barriers. Finally, the results obtained by perturbation boundary element method is compared with the analytical solution and is available for the transient pressure analysis of arbitrary shaped reservoirs.
文摘目的获取场景图像中的文本信息对理解场景内容具有重要意义,而文本检测是文本识别、理解的基础。为了解决场景文本识别中文字定位不准确的问题,本文提出了一种高效的任意形状文本检测器:非局部像素聚合网络。方法该方法使用特征金字塔增强模块和特征融合模块进行轻量级特征提取,保证了速度优势;同时引入非局部操作以增强骨干网络的特征提取能力,使其检测准确性得以提高。非局部操作是一种注意力机制,能捕捉到文本像素之间的内在关系。此外,本文设计了一种特征向量融合模块,用于融合不同尺度的特征图,使尺度多变的场景文本实例的特征表达得到增强。结果本文方法在3个场景文本数据集上与其他方法进行了比较,在速度和准确度上均表现突出。在ICDAR(International Conference on Document Analysis and Recognition)2015数据集上,本文方法比最优方法的F值提高了0.9%,检测速度达到了23.1帧/s;在CTW(Curve Text in the Wild)1500数据集上,本文方法比最优方法的F值提高了1.2%,检测速度达到了71.8帧/s;在Total-Text数据集上,本文方法比最优方法的F值提高了1.3%,检测速度达到了34.3帧/s,远远超出其他方法。结论本文方法兼顾了准确性和实时性,在准确度和速度上均达到较高水平。
