Benthic foraminiferal δ~(13)C minimum events in the southeastern Okhotsk Sea over the last 180 ka

A total of six d13C minimum events,i.e.,VI,V,IV,III,II,and I,were observed via a stable carbon and oxygen isotope analysis of infaunal benthic foraminifera Uvigerina spp.in gravity core OS03-1 in the southeastern Okhotsk Sea over the last 180 ka.These events occurred at112–109,102–90,85–76,57–54,44–40,and 17–10 ka BP.The largest negative excursions reached 2.5%in event V and were greater than 1%in the other events.We proposed that all d13C minimum events were caused by the increase in sea surface water productivity,the weakened formation of Okhotsk Sea intermediate water,and the enhancement of the oxygen minimum zone.The negative excursions were unaffected by methane hydrate destabilization and subsequent methane release based on the results obtained by using archaeal lipid markers.     

Semantic image annotation based on GMM and random walk model

Automatic image annotation has been an active topic of research in computer vision and patternrecognition for decades.A two stage automatic image annotation method based on Gaussian mixturemodel (GMM) and random walk model (abbreviated as GMM-RW) is presented.To start with,GMM fitted by the rival penalized expectation maximization (RPEM) algorithm is employed to estimatethe posterior probabilities of each annotation keyword.Subsequently, a random walk processover the constructed label similarity graph is implemented to further mine the potential correlations ofthe candidate annotations so as to capture the refining results, which plays a crucial role in semanticbased image retrieval.The contributions exhibited in this work are multifold.First, GMM is exploitedto capture the initial semantic annotations, especially the RPEM algorithm is utilized to train themodel that can determine the number of components in GMM automatically.Second, a label similaritygraph is constructed by a weighted linear combination of label similarity and visual similarity ofimages associated with the corresponding labels, which is able to avoid the phenomena of polysemyand synonym efficiently during the image annotation process.Third, the random walk is implementedover the constructed label graph to further refine the candidate set of annotations generated byGMM.Conducted experiments on the standard Corel5k demonstrate that GMM-RW is significantlymore effective than several state-of-the-arts regarding their effectiveness and efficiency in the task of automatic image annotation.     

水泥稳定沙砾试铺方案探讨

对水泥稳定沙砾试铺筑过程中的人员安排、机械配备、施工工艺及现场试验等内容进行了总结,确定了施工中的施工工艺、松铺系数、碾压遍数等。     

沥青稳定碎石ATB-30柔性基层配合比设计

ATB-30沥青稳定碎石柔性基层的配合比设计及其结构特点、油石比的确定，详细介绍了原材料的选择，得出合理的配合比指导施工。     

基于GMM和MRF的自适应阴影检测

运动目标阴影在很大程度上会影响运动目标跟踪、行为识别的正确性和有效性.为此,文中提出了一种基于混合高斯模型和马尔科夫随机场的自适应阴影检测方法.该方法首先对混合高斯模型进行改进,使其可以自适应调整参数学习率以消除浅阴影;然后采用马尔科夫随机场综合邻域的空间依赖性信息进行精确的阴影检测.为了提高基于马尔科夫随机场的阴影检...     

Response of the northwestern Pacific upper water δ~(13)C to the last deglacial ventilation of the deep Southern Ocean

The deglacial δ~13C minimum events that originated from the ventilation of the deep Southern Ocean around Antarctica, have been recorded in a range of marine sediments from the southern to tropical oceans in late Pleistocene. However, the broad δ~13C minimum event was also reported as far as to the northern middle latitudes, in northwestern Pacific marginal sea areas, during the last deglaciation. In the northwestern Pacific, forcing from the northern high latitudes is strongly expressed, while the records of in- fluence from the southern high latitudes are few. The Kuroshio Source Region (KSR) forms a boundary between the northwestern Pacific and the southern, tropical Pacific. So, high-resolution planktonic foraminiferal records in core MD06-3054 from the KSR are well positioned to identify signals from the southern hemisphere in the northwestern Pacific. Planktonic foraminiferal tests from the upper 1030 cm of the core were subject to AMS14C, carbon and oxygen isotopic measurements. A negative excursion was found to occur from about 20.0-6.0 ka BP in δ~13C records of both surface (Globigerinoides ruber) and subsurface (Pulleni- atina obliquiloculata) dwellers, but the overall trends of the two curves have reversed since 26.5 ka BP. Moreover, the δ~13C rec- ord of G. ruber (the surface dweller) shows a robust link to the record of atmospheric CO2, and its changes precede the records of P. obliquiloculata (the subsurface dweller). According to the hydrologic conditions, the broad δ~13C minimum event recorded in the KSR is also a response to the increasing ventilation of the deep Southern Ocean around Antarctica during the last deglaciation. The inconsistency between the records of the surface and subsurface dwellers was possibly caused by the ways that the low δ~13C signal was transmitted. Subsurface water primarily received the low δ~13C signal from the Antarctic Intermediate Water (AAIW), whereas the surface water was probably mainly impacted by atmospheric CO2 in the KSR. The records from the KSR confirm the deduction that the broad δ~13C minimum event in the Okinawa Trough was due to the impact of tropical Pacific surface water dur- ing the last deglaciation, and suggest that signals from the southern high latitudes also can be delivered to the northern middle latitudes.     

菌糠复合剂对不结球白菜生长发育及产量的影响

为确定菌糠复合剂对不结球白菜生长发育及产量的影响,采用盆栽试验的方法,施用不同的菌糠复合剂,定期测定了3茬不结球白菜的株高、根长、叶片数及产量。结果表明:菌糠复合剂可以促进不结球白菜生长、增加产量。其中,以含有链霉菌(Streptomycessp.)和巨大芽孢杆菌(Bacillus megatherium)的微生物组合1 菌糠处理的效果最好。     

等离子熔积成形混相瞬态场的Level-Set方法模拟

提出了一种二维等离子熔积成形瞬态模型,该模型描述了液／气界面的自由表面发展,并模拟了熔池内流体流动和传热．采用Level—Set方法处理液／气界面边界条件,考虑了熔体流动的主要驱动力——表面张力梯度、表面曲率、浮力以及工件表面的对流散热等因素．用固液相统一模型来描述固／液界面处的熔融和凝固过程,并开发了相应的软件．对高温合金K163在不同扫描速度下的熔积层表面形貌、温度场以及熔池内流场进行了模拟分析．     

异常值和未知观测噪声鲁棒的卡尔曼滤波器

给出了对异常值和未知分布的观测噪声鲁棒的卡尔曼滤波器.分析表明当以Huber损失函数替代推导卡尔曼滤波器最大后验准则中观测误差的l2范数时,就构造了一个新的准则.由于Huber损失函数可同时描述l1和l2范数,因此由这个新准则推导的卡尔曼滤波器,在具有传统卡尔曼滤波器性质的同时,也有了l1范数对异常值鲁棒的特性.而当含...     

一种模糊高斯混合说话人识别模型

为了研究模糊聚类算法在高斯混合模型（GMM）参数获取方面的应用,采用模糊C均值算法（FCM）进行语音特征矢量的聚类,并结合Tabu搜索算法得到全局最优的聚类结果,进一步用EM算法得到GMM模型参数.使用TIMIT数据库中的语音进行测试,开集和闭集说话人辨认实验都表明,该方法获取的GMM参数比普通EM算法获得的GMM模型参数性能更优,能有效降低说话人辨认系统的误识率.