首页 | 官方网站   微博 | 高级检索  
文章检索
  按 检索   检索词:      
出版年份:   被引次数:   他引次数: 提示:输入*表示无穷大
  收费全文   39篇
  免费   1篇
  国内免费   8篇
工业技术   48篇
  2023年   1篇
  2022年   4篇
  2021年   4篇
  2020年   5篇
  2019年   3篇
  2018年   4篇
  2017年   3篇
  2016年   1篇
  2015年   3篇
  2014年   3篇
  2013年   3篇
  2012年   2篇
  2011年   5篇
  2009年   1篇
  2007年   1篇
  2005年   2篇
  2004年   1篇
  2003年   1篇
  2002年   1篇
排序方式: 共有48条查询结果,搜索用时 15 毫秒
1.
基于深度学习方法的地震相智能识别技术可以大幅度减少人工操作。现有深度学习方法的网络模型只能提取单一接收域下的目标特征,难以获取地震相在剖面上的全局空间分布信息,模型对少数类地震相的边界刻画效果较差,且缺乏对预测结果可靠程度进行评估的手段。针对这些问题,提出一种用于地震相分类识别的深度学习方法:在U-Net模型的末端加入金字塔池化模块以提高模型获取全局信息的能力;采用一种融合交叉熵与Dice指数的目标函数,改善不均衡数据中少数类地震相边界的刻画问题;提出“预测信息熵”的概念用于评估地震相预测结果的不确定性。该研究方法应用于F3工区地震相预测的实验结果表明:改进深度学习方法在地震相预测中具有更高的精度和更良好的边界刻画能力;同时,预测信息熵指标也能够较好地评价预测结果的不确定性。  相似文献   
2.
中国煤系气共生成藏作用研究进展   总被引:4,自引:0,他引:4  
秦勇 《天然气工业》2018,38(4):26-36
煤系气共生共探与共采受到我国天然气行业的高度关注,勘探与开发试验已经取得了一定的成效。在分析煤系气地质条件特殊性的基础上,从煤系砂岩储层致密化机理、煤系气共生组合及成藏要素配置、煤系含气系统叠置性等方面,评述了我国煤系气共生成藏作用的研究进展。研究结果表明:(1)煤系气地质条件的特殊性表现在3个方面——煤系气赋存态和储层岩石类型具有多样性,其成藏效应与常规砂岩气有所不同;煤系沉积序列旋回性极强,气、水分布关系复杂多变;煤系砂岩储集体在广覆式泥质岩中镶嵌展布。(2)煤系富有机质特性以及烃源岩生气过程产生的有机酸,是煤系砂岩储层致密化的重要诱因,煤系砂岩气可能具有部分自生自储及吸附气的成藏特点,并可在一定程度上改善砂岩储层的物性。(3)煤系沉积特点决定了煤系气成藏效应主要取决于生烃强度、运移方式与输导体系、地层流体能量、区域有效盖层等4个方面,特殊的输导体系使得烃源岩生成的天然气在复杂的源储系统中得以重新分配,这是煤系气共生成藏的重要基础。(4)含气系统叠置性是煤系气地质研究的前缘方向之一,近年来发展了测井响应识别技术和含气系统叠置性判识方法,发现煤系存在3种典型流体压力曲线类型,并初步应用于煤系气共采有利层段的优选。  相似文献   
3.
Because mixtures of exponential distributions (with constant failure rate) have the decreasing failure rate (DFR) property, as shown by Proschan in the sixties, it is not unexpected that mixtures of distributions that have a mildly increasing failure rate (IFR) also have this property. What is. perhaps, surprising is that mixtures of IFR distributions with rapidly increasing failure rate may also behave in this manner. Two striking examples of such mixtures are presented here and illustrated graphically. This phenomenon could be the cause of concern, in practice, when it is suspected that a sample is based on pooled data.  相似文献   
4.
Coding and pooling, the major two sequential procedures in sparse coding based scene categorization systems, have drawn much attention in recent years. Yet improvements have been made for coding or pooling separately, this paper proposes a spatially constrained scheme for sparse coding on both steps. Specifically, we employ the m-nearest neighbors of a local feature in the image space to improve the consistency of coding. The benefit is that similar image features will be encoded with similar codewords, which reduced the stochasticity of a conventional coding strategy. We also show that the Viola–Jones algorithm, which is well-known in face detection, can be tailored to learning receptive fields, embedding the spatially constrained information on the pooling step. Extensive experiments on the UIUC sport event, 15 natural scenes and the Caltech 101 database suggests that scene categorization performance of several popular algorithms can be ubiquitously improved by incorporating the proposed two spatially constrained sparse coding scheme.  相似文献   
5.
针对糖尿病视网膜病变分级检测中标定样本少、多目标检测的问题,提出了一种基于改进型GoogLeNet的弱监督目标检测网络。首先,对GoogLeNet网络进行改进,去掉最后一个全连接层并保留检测目标的位置信息,添加全局最大池化层,以sigmoid交叉熵作为训练的目标函数以获得带有多种特征位置信息的特征图;然后,基于弱监督方法仅使用类别标签对网络进行训练;其次,设计一种连通区域算法来计算特征连通区域边界坐标集合;最后在待测图片中使用边界框定位病灶。实验结果表明,在小样本条件下,改进模型准确率达到了94.5%,与SSD算法相比,准确率提高了10%。改进模型实现了小样本条件下端到端的病变识别,同时该模型的高准确率保证了模型在眼底筛查中具有应用价值。  相似文献   
6.
Pooled versus separate measurements of tree-ring stable isotopes   总被引:1,自引:0,他引:1  
δ13C and δ18O of tree rings contain time integrated information about the environmental conditions weighted by seasonal growth dynamics and are well established as sources of palaeoclimatic and ecophysiological data. Annually resolved isotope chronologies are frequently produced by pooling dated growth rings from several trees prior to the isotopic analyses. This procedure has the advantage of saving time and resources, but precludes from defining the isotopic error or statistical uncertainty related to the inter-tree variability. Up to now only a few studies have compared isotope series from pooled tree rings with isotopic measurements from individual trees. We tested whether or not the δ13C and the δ18O chronologies derived from pooled and from individual tree rings display significant differences at two locations from the Iberian Peninsula to assess advantages and constraints of both methodologies. The comparisons along the period 1900-2003 reveal a good agreement between pooled chronologies and the two mean master series which were created by averaging raw individual values (Mean) or by generating a mass calibrated mean (MassC). In most of the cases, pooled chronologies show high synchronicity with averaged individual samples at interannual scale but some differences also show up especially when comparing δ18O decadal to multi-decadal variations. Moreover, differences in the first order autocorrelation among individuals may be obscured by pooling strategies. The lack of replication of pooled chronologies prevents detection of a bias due to a higher mass contribution of one sample but uncertainties associated with the analytical process itself, as sample inhomogeneity, seems to account for the observed differences.  相似文献   
7.
Cells of six strains of Cronobacter were subjected to dry stress and stored for 2.5 months at ambient temperature. The individual cell lag time distributions of recovered cells were characterized at 25 °C and 37 °C in non-selective broth. The individual cell lag times were deduced from the times taken by cultures from individual cells to reach an optical density threshold. In parallel, growth curves for each strain at high contamination levels were determined in the same growth conditions. In general, the extreme value type II distribution with a shape parameter fixed to 5 (EVIIb) was the most effective at describing the 12 observed distributions of individual cell lag times. Recently, a model for characterizing individual cell lag time distribution from population growth parameters was developed for other food-borne pathogenic bacteria such as Listeria monocytogenes. We confirmed this model’s applicability to Cronobacter by comparing the mean and the standard deviation of individual cell lag times to populational lag times observed with high initial concentration experiments. We also validated the model in realistic conditions by studying growth in powdered infant formula decimally diluted in Buffered Peptone Water, which represents the first enrichment step of the standard detection method for Cronobacter. Individual lag times and the pooling of samples significantly affect detection performances.  相似文献   
8.
文章针对传统结构Web应用系统的不足,探讨了基于Web与MIDAS的多层应用系统体系结构,在此结构下,可利用对象池实现远程数据模块重用,利用对象代理实现分布式负载均衡与容错,从而提高了系统的性能。  相似文献   
9.
骆健  蒋旻 《计算机应用》2017,37(1):255-261
针对传统的颜色-深度(RGB-D)图像物体识别的方法所存在的图像特征学习不全面、特征编码鲁棒性不够等问题,提出了基于核描述子局部约束线性编码(KD-LLC)的RGB-D图像物体识别方法。首先,在图像块间匹配核函数基础上,应用核主成分分析法提取RGB-D图像的3D形状、尺寸、边缘、颜色等多个互补性核描述子;然后,分别对它们进行LLC编码及空间池化处理以形成相应的图像编码向量;最后,把这些图像编码向量融合成具有鲁棒性、区分性的图像表示。基于RGB-D数据集的仿真实验结果表明,作为一种基于人工设计特征的RGB-D图像物体识别方法,由于所提算法综合利用深度图像和RGB图像的多方面特征,而且对传统深度核描述子的采样点选取和紧凑基向量的计算这两方面进行了改进,使得物体类别识别率达到86.8%,实体识别率达到92.7%,比其他同类方法具有更高的识别准确率。  相似文献   
10.
针对当前卷积神经网络(CNN)利用端层特征进行面部表情识别存在模型结构繁琐、训练参数过多、识别不够理想的问题,提出一种基于改进CNN与支持向量机(SVM)相结合的优化算法。首先,利用连续卷积的思想设计网络模型,以获取更多非线性激活;然后,采用自适应全局平均池化(GAP)层取代传统CNN中的全连接层,以减少网络参数量;最后,用SVM分类器代替传统Softmax函数实现表情识别,以提高模型泛化能力。实验结果表明,所提算法在Fer2013和CK+数据集上分别取得了73.4%和98.06%的识别准确率,与传统LeNet-5算法相比,在Fer2013数据集上提升了2.2个百分点,且该网络模型结构简单、参数量较少,具有良好的鲁棒性。  相似文献   
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司    京ICP备09084417号-23

京公网安备 11010802026262号