一种高准确率的交通监控视频车辆检测算法

针对高速视频监控的车辆识别易受几何变换影响且准确度不高的问题,提出了高效的阴影去除和车辆分类两种视频处理算法来提高车辆识别性能。利用阴影区域前景和背景的对比度参数进行有效的阴影去除;利用Hu不变矩表征车辆,从而克服了几何变换的影响。对比试验结果证明,本算法具有较高的阴影检测性能和极佳的车辆分类性能,明显优于其他同类型算法。     

Two New Compounds from Goniothalamus Cheliensis Hu

Goniothalamus cheliensis Hu (Annonaceace) is a small tree or shrub in Yunnan Province, southwest of China. Acetogenins, styryllactones and alkaloids1-3 from the genus were reported in the previous literature and many of them showed strong cytotoxic activities against a number of human cancer cell lines4. The titled plant has been the subject of our investigation due to the impressive cytotoxicity of its organic extract against mouse lymphocytic leukemia cells in preliminary pharmacological s…     

二维轴对称图象高阶胡氏不变矩快速算法的改进

提出了基于二维轴对称性目标图象矩快速算法的实用化改进方案.算法实用化改进是通过两个命题的证明实现的.一是通过轴对称目标图象质心在对称轴上的命题证明,实现了将二维轴对称图象矩快速算法用于轴对称性图象中心矩的计算;二是通过目标图象以坐标系中任意两点为圆心,旋转相同角度,结果图象具有平移性的命题证明,实现了算法用于对旋转不同角度的对称目标图象的中心矩的计算改进.在此基础上,完成了二维轴对称目标图象高阶胡氏不变矩实用化快速算法的实现.实验证明,该算法具有较好的实时性能,且具有较小的引入误差.     

结合区域提取和改进卷积神经网络的水下小目标检测

针对水下小目标信息量有限而难以提取有效特征导致的检测性能不佳问题,提出了一种结合区域提取和融合Hu矩特征的改进卷积神经网络水下小目标检测方法。该方法包含区域提取和分类两个步骤。首先以马尔可夫随机场分割算法为基础进行区域提取,对潜在目标定位的同时降低伪目标对后续分类的干扰;然后提取潜在目标区域的Hu矩特征并融入卷积神经网络,形成一种形状特征表征能力更强的改进卷积神经网络用于分类。声呐实测数据处理结果表明,该方法可以有效提升对水下小目标的发现概率和正确报警率,与其他目标检测方法相比,该方法具有更好的检测性能和泛化性。     

Separation and identification of multiple constituents in Xiao Chai Hu Decoction (Sho‐saiko‐to) by bioactivity‐guided fractionation combined with LC‐ESI‐QTOFMS/MS

Xiao Chai Hu Decoction (XCHD), named Sho‐saiko‐to in Japanese, is a well‐known traditional Chinese medicine formula used in Asia. However, the characterization methods used in the past have lacked sensitivity and the nature of the active constituents of XCHD remains unclear. This study was carried out to establish the hyphenated method of bioactivity‐guided fractionation and liquid chromatography coupled with electrospray ionization quadrupole time‐of‐flight mass spectrometry (LC‐ESI‐QTOFMS/MS) in order to identify the major bioactive constituents of XCHD. D101 macroporous resin was used to separate and enrich the material base into four fractions, XCHD‐1, XCHD‐2, XCHD‐3 and XCHD‐4. Each fraction was then evaluated for its antidepressant effect using depression‐related parameters. An LC‐ESI‐QTOFMS/MS method in both positive and negative ion mode was also applied for separation and identification of the biological active fractions of XCHD. As a result, 79 compounds including polysaccharides, flavonoids, saikosaponins, ginsenosides, licoricesaponins and gingerols were detected, 69 of them were identified or tentatively characterized. Based on our preliminary characterization investigations, polysaccharides, gingerols and flavonoids in XCHD may contribute to the antidepressant effect of XCHD. In conclusion, the hyphenated method of bioactivity‐guided fractionation and LC‐ESI‐QTOFMS/MS was meaningful for the isolation and preliminary identification of the biological active components in complex matrices of traditional Chinese medicine. Copyright © 2014 John Wiley & Sons, Ltd.     

Highly Oxygenated Triterpenoids and Diterpenoids from Fructus Rubi (Rubus chingii Hu) and Their NF-kappa B Inhibitory Effects

During a phytochemical investigation of the unripe fruits of Rubus chingii Hu (i.e., Fructus Rubi, a traditional Chinese medicine named “Fu-Pen-Zi”), a number of highly oxygenated terpenoids were isolated and characterized. These included nine ursane-type (1, 2, and 4–10), five oleanane-type (3, 11–14), and six cucurbitane-type (15–20) triterpenoids, together with five ent-kaurane-type diterpenoids (21–25). Among them, (4R,5R,8R,9R,10R,14S,17S,18S,19R,20R)-2,19α,23-trihydroxy-3-oxo-urs-1,12-dien-28-oic acid (rubusacid A, 1), (2R*,4S*,5R*,8R*,9R*,10R*,14S*,17S*, 18S*,19R*,20R*)-2α,19α,24-trihydroxy-3-oxo-urs-12-en-28-oic acid (rubusacid B, 2), (5R,8R,9R,10R, 14S,17R,18S,19S)-2,19α-dihydroxy-olean-1,12-dien-28-oic acid (rubusacid C, 3), and (3S,5S,8S,9R, 10S,13R,16R)-3α,16α,17-trihydroxy-ent-kaur-2-one (rubusone, 21) were previously undescribed. Their chemical structures and absolute configurations were elucidated on the basis of spectroscopic data and electronic circular dichroism (ECD) analyses. Compounds 1 and 3 are rare naturally occurring pentacyclic triterpenoids featuring a special α,β-unsaturated keto-enol (diosphenol) unit in ring A. Cucurbitacin B (15), cucurbitacin D (16), and 3α,16α,20(R),25-tetrahydroxy-cucurbita-5,23- dien-2,11,22-trione (17) were found to have remarkable inhibitory effects against NF-κB, with IC₅₀ values of 0.08, 0.61, and 1.60 μM, respectively.     

红外激光主动成像和识别

在传统激光主动成像系统的基础上,结合目标识别技术搭建了一台激光主动成像识别系统实验平台,研究了激光主动成像后的目标识别技术。由7个不变Hu矩构成特征量,用由136个权值系数构成BP神经网络算法对黑夜条件下450m处的运动目标—43式冲锋模具枪进行了实验研究。研究显示,采用该方法成功获得了清晰的红外激光主动成像效果,对2740frame激光主动成像图像的统计目标识别率达到了68．87％,其中旋转变换下的统计识别率可达80．05％。该项研究对实际黑夜暗小目标的探测识别具有重要意义。     

柳叶蜡梅叶挥发油化学成分GC-MS分析

分析鉴定柳叶蜡梅叶挥发油化学成分,为其全面开发利用提供科学依据。采用了水蒸气蒸馏法提取2批柳叶蜡梅叶中挥发性化学成分,采用气相色谱-质谱(GC-MS1)联用技术对其化学成分进行了分离鉴定。2批样品均分离了29个峰并鉴定出其中25个组分,用面积归一化法测得其相对质量百分含量     

The antioxidant compounds isolated from the fruits of chinese wild raspberry Rubus Chingii Hu

Abstract

Raspberry, the fruit of Rubus Chingji Hu, is a widely distributed economic staple food in China. It has long been used as a traditional medicine in mainland China to treat kidney enuresis, nocturnal emission and premature ejaculation in clinic. In this paper, six known compounds (1?–?6) were purified from the fruits of Rubus chingji. Their structures were elucidated as (16α)-16,17-dihydroxy-ent-kauran-2-one17-O-β-D-glucopyranoside (1), (16R) -16,17-dihydroxy-ent-kaurane-2-one (2), 3,3′-di-O-methylellagic acid 4-(5′′-acetyl)-α-L-arabinofuranoside (3), quercilicoside A (4), esculetine (5) and ethyl-β-D-glucoside (6). All the compounds were isolated from Rubus Chingji Hu for the first time. Compounds 3 and 5 shown distinctive free radical scavenging activities in DPPH and FRAP assays. In addition, no cytotoxicity was observed for compounds 3 and 5 against different cancer cells, suggesting that they might be useful as potential antioxidant agents against various reactive oxygen species.