以一种新的距离度量作为反馈的相关跟踪方法

基于相关的模板匹配方法在图像跟踪中得到了广泛的应用,但是这种方法的主要缺点是对图像噪声非常敏感,而且由这种方法得出的相关曲面比较平坦。为了有效抑制噪声和局部遮挡对图像的影响,提出用鲁棒统计中的一个值代替相关值作为距离度量。它不是计算两幅图像的对应像素差值,而是计算两幅图像中差别不大的对应像素对的个数。实验证明这样得到的曲面相对　　收稿日期:2003 08 31;　修订日期:2003 10 28作者简介:张桂林(1944 ),男,天津人,教授,博士生导师,主要研究方向为自动识别、人体生物统计特征识别、图像处理算法与系统性能评估等。　　　　　　　要尖锐得多。当上述的距离度量应用于跟踪过程中时,模板的合理更新对跟踪非常重要。基于上述的距离度量方法,提出了一种新的模板更新策略。利用两幅图像中近似像素在整修匹配模板像素中所占的比例大小决定加权系数。这种加权策略可以自适应地根据图像序列的变化选择更新相关模板,同时在一定程度上减小模板漂移的危险。     

注凝析油开采稠油的地质工艺

在西西伯利亚俄罗斯油田开展了注凝析油开发稠油的研究，现场实验结果表明，注入25%的凝析油，可使其混合液粘度达到原高粘油粘度的1/60，以俄罗斯油田为例，探讨了注凝析油矿场试验的可行性，预测了凝析油的用量和可能产出的油量，并对原油外输提出了一系列更换工艺设备的要求。     

基于奇异值分解的特征跟踪方法

在传统的基于模板匹配的跟踪方法中，均是给定一个模板，然后从图像中各个位置取出一个个与模板大小一致的区域进行相似性度量，找出与模板距离最小的一个区域作为当前模板，以便进行下一步的匹配跟踪工作。在景象匹配和相关跟踪过程中，由于所面临的大多数是变化的场景，实时获取的图像与预存模板之间存在比较大的差异，传统相关匹配方法的应用就会受到限制；而且在跟踪过程中，随时更新模板会造成跟踪性能对扰动过分敏感，从而产生漂移。首先拍摄目标不同角度的图像（尽可能包含目标可能出现的所有情况），构成目标图像训练集合，抽取出特征矩阵，对它进行奇异值分解，构成一个关于目标的多维空间。然后再用匹配方法在全局范围搜索，找出目标的大致位置，并利用收敛方法在确定的大致位置内进行搜索，确定目标的仿射变换系数，从而得到一个目标位置的确切描述。     

基于OSEK规范的网络管理的研究与改进

首先论述了基于OSEK规范的网络管理特点并对OSEK网络管理的工作机制进行了详细分析;在基于OSEK网络管理特点和工作机制的基础上,对OSEK网络管理在以太网上的扩展、资源占有率优化和实时性提高几个方面给出了解决方案。     

基于SPW的流星余迹极低信噪比通信系统建模与仿真*

SPW具有丰富的通信库和DSP库,且具有灵活的自定义模块能力,是一种复杂通信系统建模和仿真工具.流星余迹通信是一种特殊的通信方式,在通信领域中具有特殊的重要地位.介绍了流星余迹极低信噪比通信系统原理的基础,给出了基于SPW的流星余迹极低信噪比通信系统仿真模型及仿真结果.     

In vitro comparison of hepatic metabolism of 9cis-11 trans and 10trans-12cis isomers of CLA in the rat

Hepatic metabolism of the two main isomers of CLA (9cis-11 trans, 10trans-12cisC_18∶2) was compared to that of oleic acid (representative of the main plasma FA) in 16 rats by using the in vitro method of incubated liver slices. Liver tissue samples were incubated at 37°C for 17h under an atmosphere of 95% O₂/5%CO₂ in a medium supplemented with 0.75 mM of FA mixture (representative of circulating nonesterified FA) and with 55 μM [1-¹⁴C]9cis-11 trans C_18∶2, [1-¹⁴C]10trans-12cis C_18∶2, or [1-¹⁴C]oleate. The uptake of CLA by hepatocytes was similar for both isomers (9%) and was three times higher (P<0.01) than for oleate (2.6%). The rate of CLA isomer oxidation was two times higher (49 and 40% of incorporated amounts of 9cis-11 trans and 10trans-12 cis, respectively) than that of oleate (P<0.01). Total oxidation of oleate and CLA isomers into [¹⁴CO₂] was low (2 to 7% of total oxidized FA) compared to the partial oxidation (93 to 98%) leading to the production of [¹⁴C] acid-soluble products. CLA isoemrs escaping from catabolism were both highly desaturated (26.7 and 26.8%) into conjugated 18∶3. Oleate and CLA isomers were mainly esterified into neutral lipids (30%). They were slowly secreted as parts of VLDL particles (<0.4% of FA incorporated into cells), the extent of secretion of oleate and of 10trans-12 cis being 2.2-fold higher than that of 9cis-11 trans (P<0.02). In conclusion, this study clearly showed that both CLA isomers were highly catabolized by hepatocytes, reducing their availability for peripheral tissues. Moreover, more than 25% of CLA escaping from catabolism was converted into conjugated 18∶3, the biological properties of which remain to be elucidated.     

Effects of conjugated linoleic acid isomers on the hepatic microsomal desaturation activities in vitro

The influence of individual conjugated linoleic acid (CLA) isomers on the Δ6 desaturation of linoleic and α-linolenic acids and on the Δ9 desaturation of stearic acid was investigated in vitro, using rat liver microsomes. The Δ6 desaturation of 18∶2n−6 was decreased from 23 to 38% when the ratio of 9cis,11trans-18∶2 to 18∶2n−6 increased from 0.5 to 2. The compound 10trans,12cis-18∶2 exhibited a similar effect only at the highest concentration. The Δ6 desaturation of α-linolenic acid was slightly affected by the presence of CLA isomers. The sole isomer to induce an inhibitory effect on the Δ9 desaturation of stearic acid was 10trans,12cis-18∶2.     

遗传算法理论及其应用研究进展

首先阐述遗传算法的原理和求解问题的一般过程;然后讨论了近年来从遗传算子、控制参数等方面对遗传算法的改进,并对遗传算法在计算机科学与人工智能、自动控制以及组合优化等领域的应用进行陈述;最后评述了遗传算法未来的研究方向和主要研究内容。     

基于GPS实时数据的在线过滤与补遗研究*

为提高GPS数据的有效性与可靠性,使之为实现实时交通流量预测与交通诱导服务,首先基于GPS历史数据,以变异系数极小化为优化目标进行GPS数据过滤模型的优选,并对实时数据缺损提出两种补遗算法;然后给出基于动态GPS实时数据的在线数据过滤与补遗一体化算法;最后结合2008年杭州市GPS历史数据对模型进行应用。结果表明,基于简单算术平均的滤波模型即为GPS数据过滤的最佳模型;基于时间序列缺失数据的快速补遗算法能够很好地满足实时预报要求,实现快速补遗。