铝合金TIG焊熔池正面图像模式识别

将图像处理与模式识别技术用于铝合金TIG焊接过程信息提取过程，根据铝合金熔池图像随机噪声强的特点，采用加权中值滤波，统计灰度边缘检测、统计期望阈值法和投影方法对铝合金熔池图像进行了预处理。探索了将神经网络用于焊接熔池图像处理的方法，采用BP神经网络对二值化熔池图像进行边缘提取，取得了理想的效果。研究了大电流条件下铝合金熔池图像的对称性，通过单面图像，得到了完全的熔池边缘图像。     

EPS模陶瓷型精密铸造失模工艺研究

针对EPS陶瓷壳型精铸失模时容易裂壳的现象,对EPS消失模精密铸造工艺进行了研究。利用自制的膨胀率检测设备对不同密度的EPS模试样在不同加热温度下的最大膨胀率进行了测试,研究了壳型膨胀与失模温度之间的关系;对陶瓷壳型中EPS模失模时的温度场分布进行测量,发现EPS模试样的密度大小对失模时温度场有着很重要的影响。在试验和分析的基础上,对EPS模失模机制进行了探讨,提出了高温失模机制。高温失模有利于降低失模时对壳型的膨胀作用力,为获得洁净而健全的陶瓷型壳提供了可靠保证。     

基于电容层析成像和模糊模式识别的油气两相流流型辨识新方法的研究

基于电容层析成像和模糊模式识别技术别提出了一种油气两相流流型辨识的新方法。建立了12电极电容层析成像流型自动识别系统，该系统利用Tikhonov正则化原理并结合SIRT(Simultaneous Reconstruction Techniques)算法进行图像重建。Tikhonov正则化原理用于克服图像重建过程中的不适定问题，SIRT算法用于提高最终重建图像的质量。根据流型的随机和模糊特性，提出了一种根据管截面重建图像进行流型辨识的模糊流型判别方法。研究结果表明，提出的流型辨识新方法是有效的。对于层状流、核心流、环状流、均相流等流型，流型辨识的准确率高于95％，辨识一个流型所用的时间小于0.3秒。对于塞状流，流型辨识的准确率高于90％。     

我国天然柠檬醛的利用及鸢尾酮的合成进展

我国富含柠檬醛的山苍子油资源丰富 ,每年出口约 2 0 0 0t,在世界香料市场占有特殊的地位。为了改变该资源低值单调的利用途径 ,系统地总结和评述了山苍子油半合成高档香料———鸢尾酮以及多方面综合利用的研究进展 ,阐明了将我国天然柠檬醛的资源优势转化为可持续发展的经济优势的重要意义 ,提出了以栽培种植为基础、以半合成鸢尾酮系列香料为主导、以环境友好和高效综合利用为目的的产业化模式     

Fingerprinting G-protein-coupled receptors

Recently we reported the design of a discriminating fingerprintfor rhodopsin-like G-protein-coupled receptors (GPCRs). Thefingerprint encodes the seven putative membrane-spanning motifsand was potently diagnostic of all GPCRs (52 in all) in version8.1 of the OWL composite sequence database, readily distinguishingthem from all other integral membrane proteins. With a 3-foldincrease in the size of OWL, the fingerprint has been updatedand now finds 332 receptors that match all the motifs. The situation,however, has grown in complexity: 61 sequences make imperfectmatches with the fingerprint, yielding a total of 393 ‘hits’.The bulk of the partial hits are olfactory receptors: theseappear to fall into discrete subfamilies in which one or moreof the transmembrane motifs are either poorly matched or arenot matched at all. These results are supported by preliminaryphylogenetic analyses, which show the olfactory and variousother partial matches clustering away from the main body oftrue hits. The approach has provided a powerful diagnostic toolfor identifying GPCRs, and results are consistent with previousobservations that the pheromone, cAMP and secretin-like receptorsbelong to separate families-these bear their own unique sequencefingerprints by which they may be distinguished from the rhodopsin-likesuperfamily     

PARTICLE FLOW PATTERNS IN THE RISER AND DOWNER OF CIRCULATING FLUIDIZED BED

This paper presents an experimental study on the flow patterns of FCC particles in a 140 mm ID Circulating Fluidized Bed with concurrent upflow and downflow gas-solid suspension. Based on the distribution of local particle velocity and particle concentration measured by a Fiber-Optical Probe Laser Doppler Velocimeter and a Fiber Optical Probe System respectively, the different flow patterns of local particls concentration, local particle velocity, local particle fluctuating velocity and sectionally average particle velocity in concurrent upflow and downflow gas-solid system have been investigated. It is found that the particle flow in the concurrent downflow is much more uniform radially than that in the concurrent upflow riser. The investigation of flow patterns in different flow systems is of significance to the development of a new gas-solid reactor.     

基于交叉皮质模型的单幅图像阴影检测算法

图像中含有阴影区域对后续处理任务影响较大, 根据阴影特性, 提出基于交叉皮质模型(Intersecting cortical model, ICM)的单幅图像阴影检测算法. 通过在点火连接矩阵构造上考虑邻域像素值依赖关系, 融入局部二值模式(Local binary pattern, LBP)表征的纹理信息形成了Te-ICM模型. 根据阴影检测流程,利用模型迭代特性, 通过设计停止条件自动检测本影, 在本影修复后生成附着半影. 同时优化模型参数, 设计了基于分层聚类直方图划分的阈值下降策略. 仿真结果表明: 对于典型影像集, Te-ICM模型及相应参数设计可以较好地实现阴影检测, 输出阴影掩模准确度高, 为后续阴影去除提供了基础.     

案例推理属性权重的分配模型比较研究

案例推理系统中各属性权重的赋值决定了案例之间的相似度 大小,进而对推理结果的正确与否产生显著影响.以属性加权K-最近邻 相似案例检索为基础,讨论了使用注水原理分配属性权重的机理,并通过建 立权重分配的合理性指标,构造拉格朗日函数对权重进行优 化求解,得到了收敛的注水分配算法.通过五折交叉的模式分类实验 ,分别对属性权重的平均分配法、注水分配算法和遗传算法分配法进行了比较研究,案例推理分类结果证明,在引入注水分配算法后,其分类性能得到有效改善.     

Dynamic thread mapping of shared memory applications by exploiting cache coherence protocols

In current computer architectures, the communication performance between threads varies depending on the memory hierarchy. This performance difference must be considered when mapping parallel applications to processor cores. In parallel applications based on the shared memory paradigm, the communication is difficult to detect because it is implicit. Furthermore, dynamic mapping introduces several challenges, since it needs to find a suitable mapping and migrate the threads with a low overhead during the execution of the application. We propose a mechanism to detect the communication pattern of shared memory applications by monitoring cache coherence protocols. We also propose heuristics that, combined with our communication detection mechanism, allow the mapping to be performed dynamically by the operating system. Experiments with the NAS Parallel Benchmarks showed a reduction of up to 13.9% of the execution time, 30.5% of the cache misses and 39.4% of the number of invalidation messages.