基于人工蜂群算法的粒子滤波目标跟踪

针对基于粒子滤波的视频目标跟踪算法中由于粒子重采样过程而导致粒子贫化的问题,提出了一种基于人工蜂群算法的粒子滤波目标跟踪算法,利用群体智能的特点使得粒子集在重采样前得到优化,保持了粒子的多样性,从而解决了粒子贫化问题,同时增加了有效粒子的数目.实验结果表明,基于人工蜂群算法的粒子滤波跟踪算法,比标准粒子滤波跟踪算法所需粒子数更少,对目标遮挡、较复杂背景有较好的跟踪效果.     

基于紫外光谱法和有监督模式识别技术对氯化钠、氢氧化钠和β-苯乙胺种类的鉴别

β-苯乙胺(PEA)是一种重要的有机合成中间体,在PEA生产过程中,最终的生成物中常常为含有氯化钠(NaCl)、氢氧化钠(NaOH)和PEA三者的混合物。因此,对NaCl,NaOH,PEA和混合物种类进行鉴别,有利于PEA的合成以及定性检测,需建立紫外光谱快速鉴别NaCl,NaOH,PEA和混合物种类的方法。利用紫外光谱法分别测量了NaCl,NaOH,PEA和混合物在190～400 nm范围的吸收光谱。首先,采用偏最小二乘法(PLS)提取紫外光谱的主成分信息,用少数的主成分信息取代原始变量,减少模型的复杂度。用PLS提取NaCl,NaOH和PEA在前三个主成分空间中得分向量值的分布。可知NaCl,NaOH和PEA前三个主成分累计贡献率分别是96.64%,99.44%和99.95%。因此,NaCl、NaOH和PEA的前三个主成分基本包含了大部分的光谱信息。其次,用获得的三个主成分作为输入变量,采用线性判别分析（LDA）,Sigmoid SVM,RBF-SVM,RBF-ANN,BP-ANN和人工蜂群(ABC)优化神经网络(ABC-BP-ANN)等模式识别方法对NaCl,NaOH和PEA种类进行判别,获得总的敏感性分别为95.6%,95.6%,95.9%,95.8%,96.9%和99.6%。由于NaCl和NaOH特征吸收峰很相似,主成分得分向量会出现重叠现象,导致NaCl和NaOH的种类鉴别出现误判。通过对比六种分类方法,可知ABC-BP-ANN效果最优,BP-ANN次之,RBF-SVM和RBF-ANN结果相似,但比BP-ANN稍差,LDA和Sigmoid-SVM效果最差。最后,配制7种不同摩尔分数的混合物(混合物摩尔分数是指PEA物质的量占混合物总物质的量百分比),浓度范围为0%mol·L-1～60%mol·L-1,然后采用RBF-SVM,BP-ANN和ABC-BP-ANN三种方法对混合物种类进行判别。从敏感性和特异性结果可以得知,ABC-BP-ANN分类效果最好,BP-ANN次之,RBF-SVM分类效果最差, 由混合物得到的结果与单组分的结果相一致。结果表明,紫外光谱结合ABC-BP-ANN模式识别方法可以成功区分NaCl,NaOH,PEA和混合物的种类。该方法可作为一种简便、快速、可靠的方法用于NaCl,NaOH,PEA和混合物的种类判别,并为PEA的合成和质量控制提供理论依据和技术支撑。     

Enhancement of wound healing via topical application of natural products: In vitro and in vivo evaluations

Intact skin is the first physical barrier against all microbial infections. Thus, in the cases of wounds, burns, and skin damage, bacteria can infect and invade the deeper layers of skin to the bloodstream and other organs leading to severe illnesses. Thus, our study aims to investigate the potential activity of natural products, propolis and honeybee venom, to control wound infections with multi-drug resistant Staphylococcus aureus (MDRSA) and safely accelerate the wound healing. First, this study characterized the clinically isolated S. aureus using biochemical, molecular, and antibiotic sensitivity tests. Then, the hydrogel was prepared via mixing chitosan with honey, propolis, and venom at different ratios, followed by physicochemical characterization and biological examination. The in vivo experiment results after topical application of optimum concentrations revealed that both venom and propolis have significant antibacterial activity at different temperatures. The IC₅₀ of both propolis antioxidant and cytotoxicity assays was found to be 40.07 ± 2.18 μg/mL and 18.3 μg/mL, respectively. The cocktail bacteria showed both minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC) of 10 µg/mL and of 300 µg/mL with venom respectively & MIC and MBC of 100 µg/mL, 300 µg/mL with propolis respectively. The use of hydrogel was effective against wound infection and enhanced wound healing during 14 days. Before starting clinical trials, further studies can be done on large animal models.     

Improvement of X-ray fluorescence sensitivity by dry ashing method for elemental analysis of bee honey

Elements, K, Ca, Ti, Cr, Mn, Fe, Ni, Cu, Zn, Rb, and Sr in bee honey samples were determined using an improved dry ashing (DA) method for XRF with Mo-secondary target (Mo-XRF). The sensitivity of the DA method was significantly improved in comparison to the wet ashing (WA) and the direct (D) methods. The limits of detection (LODs) obtained by DA (3.4–0.007 µg/g) method were better by an order of magnitude than those obtained by WA (34.0–0.120 µg/g) and D (61.2–0.270 µg/g) methods. Further improvements in the sensitivity of the DA-XRF were achieved by using a Cu-secondary target for the excitation of the elements of K, Ca, Ti, Cr, and Mn. In this instance, the LODs were in the range of 0.220–0.024 µg/g. The results of DA-XRF analysis revealed a very good accuracy with errors less than 7.1% and a precision with a relative standard deviation (RSD) better than ± 8.8%.The improved DA-XRF analysis was applied for the determination of the above mentioned elements in several Syrian bee honey samples. The results were comparable to those obtained by the atomic spectrometry method with correlation coefficients better than 0.9927.     

人工蜂群算法综述

讨论人工蜂群算法（ABC, Artificial Bee Colony Algorithm）的开发和应用。首先回顾从2005年以来的开发过程，然后介绍算法的标准版本以及常用的测试函数，最后讨论该算法的一些非标准的改进、应用以及未来潜在的应用领域。     

Spectra and structure of phosphorus-boron compounds : IV. Vibrational spectra of solid trifluorophosphineborane

The Raman spectra of F₃PBH₃ and F₃PBD₃ have been recorded (2500-10 cm⁻¹) of the liquids (−80°C) and solids (−196°C) as well as the infrared spectra (4000-33 cm⁻¹) of the solids. In the spectrum of the solid state many of the ¹⁰B and ¹¹B fundamentals were clearly defined and it was also possible to assign the BH₃ torsional frequency from the infrared and Raman spectra of the solids. A complete vibrational assignment is proposed and a normal coordinate calculation carried out. The force constant of 2.46 mdyn Å⁻¹ for the P-B stretching mode is consistent with the short P-B bond; this constant is compared to the similar quantity for several other phosphorus-boron compounds. All of the E modes for the “free” molecule are shown to be split by the site symmetry which indicates that the molecules occupy C_s or C₁ sites. The large number of observed lattice modes is consistent with two or more molecules per primitive cell. The torsional frequency was observed at 224 cm⁻¹ and 167 cm⁻¹ in hydrogen and deuterium compounds in the solid, respectively. These frequencies gave a periodic barrier of 4.15 kcal mole⁻¹ for F₃PBH₃ and 4.31 kcal mole⁻¹ for F₃PBD₃. CNDO/2 calculations have been carried out for F₃PBH₃ and the isoelectronic F₃SiCH₃ molecule in both the staggered and eclipsed forms and the dipole and barrier origins are discussed.     

基于嵌入式GPRS技术的大区域火灾监控系统

针对当前的大区域内火情监控以人工为主,实时性较差,报警存在延时的问题;在需求分析的基础上,设计并实现了一种基于嵌入式与GPRS技术的大区域火情监控系统,硬件平台以嵌入式 S3C2410 芯片配合温度传感器LM75A以及相关外围电路,采集处理传感温度信息,利用GPRS网络的大范围数据通信能力传递各种火情信息,软件采用了分层的设计思想,将整个系统的软件设计分为用户应用层、网络集成层、硬件驱动层3个层次结构进行设计,利用SMS的方式进行火灾报警,避免了人工检测时主观强,实时性差的弊端;系统经过大量的实际数据测试证明,误差率基本能够保持在15%以内,比传统算法,报警准确率提高了27.4%,大区域控制范围在1平方公里,报警性能良好,具有很强的实际应用价值。     

Determination of spinosad at trace levels in bee pollen and beeswax with solid–liquid extraction and LC–ESI‐MS

This paper reports the use of a new LC method with a fused‐core analytical column coupled to ESI‐MS to determine residues of the biopesticide spinosad in bee pollen and beeswax. The method analyzes the active ingredients, spinosyns A and D, with a simple and efficient sample treatment (recovery between 90 and 105%) consisting of a solid–liquid extraction with acetone (bee pollen) or acetonitrile (beeswax). The method was validated in terms of selectivity, LOD, LOQ, linearity, and precision. The LOD and LOQ values ranged between 0.1–0.2 and 0.4–0.7 μg/kg, respectively. Moreover, the precision obtained within the linear concentration range (LOQ 500 μg/kg) was satisfactory (RSD lower than 5%). Finally, the proposed method was applied to analyze bee pollen and beeswax samples collected from apiaries located close to fruit orchards in two Spanish regions.     

A novel peptide from Apis mellifera and solid-phase synthesis of its analogue

A novel peptide designated secapin-1,was purified and characterized from Apis mellifera.The molecular weight of 25 amino acid peptide secapin-1 was found to be 2821.5625 Da by ESI-FTICR-MS.It showed high identity to secapin.The sequence of secapin-1 was determined to be YIINVPPRCPPGSKFVKNKCRVIVP by automatic Edman degradation.A disulfide bond was formed between Cys⁹ and Cys²⁰ residues.In addition,an analogue of secapin-1 was synthesized by solid phase peptide synthesis method.The synthesis product was successfully purified and identified to homogeneity by using a combination of SEC,IEC,and RP-HPLC techniques.