基于多特征的快速红外弱小目标检测算法

为了从全向红外搜索和跟踪系统采集的海量大视场高分辨率红外图像中快速准确地检测出红外弱小目标,本文提出了一种基于由粗到细的分阶段检测策略和时空域特征融合的红外弱小目标检测算法.首先,通过引入基于频域的快速显著性检测算法预先检测出目标可能存在的候选区域;其次,对候选区域进行角点检测以判定是否存在候选目标;最后,通过结合帧间时空域特征对候选目标进行进一步判定,以提取真实目标、删除虚假目标.多种实际场景的实验结果表明,该目标检测算法不仅运算量小而且探测概率高、虚警率低,是一种工程实用性能很好的红外弱小目标检测算法.     

激光投影成像式运动目标位姿测量与误差分析

为了实现室内运动目标位姿的高精度测量,建立了一套激光投影成像式位姿测量系统.该系统利用两两共线且交叉排列在同一平面上的点激光投射器作为合作目标捷联在运动目标上,通过与光斑接收幕墙的配合共同组成运动目标位姿测量基线放大系统,利用高速摄像机实时记录幕墙上投影光斑的位置,利用摄像机标定结果求解投影光斑的世界坐标,利用投影光斑之间构成的单位向量建立运动目标位姿解算模型.最后,根据测量原理推导了图像坐标提取、摄像机外部参数标定、光束直线度与目标位姿解算结果之间的误差传递函数.实验结果表明,当摄像机的视场范围为14 000mm×7 000mm时,测量系统的姿态角测量精度为1′(1δ),位置测量精度为5mm,且误差大小与目标位姿测量误差传递函数理论计算值一致,验证了本文提出的目标位姿测量方法与测量误差传递模型的准确性,能够满足目标位姿测量高精度的要求.     

具有荧光探针的靶向肽/聚(醚-酰胺)制备及其肝癌细胞靶向性能研究

通过端氨基聚乙二醇PEG(Ⅰ)与二亚乙基三胺五乙酸二酐(DTPAA)开环合成新型端氨基聚(醚-酰胺)(PEG/DTPA)共聚物造影剂配体(Ⅱ)(Step1);Ⅱ的端氨基与偶联剂3-马来酰亚胺苯甲酸-N-琥珀酰亚胺酯(MBS)的活化端COOH反应,生成偶联剂/聚(醚-酰胺)MB/PEG/DTPA(Ⅲ′)(Step2);再通过Ⅲ′中MBS的CC双键与肝癌细胞靶向黏附肽FAM-AGKGTPSLETTPC-(SH)-COOH(FAM-13)上的巯基SH发生Michael加成反应(Step3),合成含有荧光探针FAM(5-carboxyfluorescein)的肝癌靶向肽/聚(醚-酰胺)(FAM-13/PEG/DTPA,Ⅲ).用1H-NMR和13C-NMR等方法对共聚物进行表征.Ⅲ对正常肝细胞L-02几乎观察不到荧光现象,而对肝癌细胞BEL-7404则有很强的黄绿色荧光,Ⅲ对肝癌细胞有很强的靶向性.大分子配体Ⅲ可望用于制备大分子造影剂及靶向载体负载药物.     

丹参活血化瘀活性成分的靶标

丹参活血化瘀活性成分主要有丹参素、迷迭香酸、丹参酮等化合物,这些化合物具有某些共同的化学结构特征.运用靶分子信息学平台,实现丹参共同化学结构特征与已知结构的蛋白质的结合能力的系统化研究,并进一步探索丹参活血作用的机理.结果表明,丹参活血化瘀活性成分主要通过调控人体内肾上腺素类激素的代谢网络,在一定程度上抑制肾上腺素类激素的分解代谢,形成积累,从而对心血管疾病具有治疗效果.丹参活血化瘀活性成分的主要靶标是苯乙醇胺N-甲基转移酶、儿茶酚O-甲基转移酶和单胺氧化酶A.     

一种用于“蛇形”机动目标的跟踪方法研究

为了满足光电跟踪系统中对“蛇形”机动目标跟踪的要求,针对以前应用的Singer模型的缺陷,根据“蛇形”机动目标运动规律,提出一种机动目标转弯模型,结合扩展卡尔曼滤波算法,对典型“蛇形”机动目标进行跟踪。通过Monte Carlo仿真,实现了在平面内对“蛇形”机动目标位置的准确跟踪,并且距离均方根误差（RMSE）比Singer模型要小很多,从而验证了该方法的有效性。     

Profiling the constituents of Dachuanxiong decoction by liquid chromatography with high‐resolution tandem mass spectrometry using target and nontarget data mining

Comprehensive characterization of the large number of compounds existing in traditional Chinese medicines is still a great challenge. In this study, a strategy of precursor ion selected acquisition coupled with target and nontarget data mining was established to systematically characterize the chemical constituents of traditional Chinese medicines. This strategy consisted of four steps: (1) precursor ion selected acquisition was developed to trigger additional tandem mass spectrometry fragmentation reactions, especially for trace constituents; (2) in‐house database of compounds was established and diagnostic characteristics were summarized; (3) compounds were identified by target and nontarget data mining; and (4) compound structures were elucidated based on accurate mass matching and comparison of fragment ions, and isomers were discriminated by the intensity of fragment ions, fragmentation pattern analysis, and calculated log P values. This strategy was successfully applied to comprehensively identify the constituents in Dachuanxiong decoction. Finally, a total of 218 compounds assigned to six categories were characterized, and 107 compounds were characterized by nontarget analysis for the first time. In addition, three new diagnostic characteristics of esters of citric acids were elucidated. This research enriched the material basis of Dachuanxiong decoction and provided a new strategy for identifying the chemical constituents of other traditional Chinese medicines.     

A target‐group‐change couple with mass defect filtering strategy to identify the metabolites of “Dogel ebs” in rats plasma,urine and bile

“Dogel ebs” was known as Sophora flavescens Ait., a classical traditional Chinese Mongolian herbal medicine, which had the effects on damp‐heat dysentery, scrofula, and syndrome of accumulated dampness toxicity. Although the chemical constituents have been clarified by our previous studies, the metabolic transformation of “Dogel ebs” in vivo was still unclear. To explore the mechanism of “Dogel ebs,” the metabolites in plasma, bile, and urine samples were investigated. A fast positive and negative ion switching technology was used for the simultaneous determination of flavonoids and alkaloids in “Dogel ebs” in a single run. And a target‐group‐change coupled with mass defect filtering strategy was utilized to analyze the collected data. 89 parent compounds and 82 metabolites were characterized by high‐performance liquid chromatography with quadrupole exactive Orbitrap mass spectrometry. Both phase I and phase II metabolites were observed and the metabolic pathways involved in oxidation, demethylation, acetylation, and glucuronidation. 69 metabolites of “Dogel ebs,” including three hydroxyls bonding xanthohumol, formononetin‐7‐O‐glucuronide, 2′‐hydroxyl‐isoxanthohumol decarboxylation metabolite, oxysophocarpine dehydrogen, 9α‐hydroxysophoramine‐O‐glucuronide, etc. were reported for the first time.     

低信噪比条件下多节点声呐目标跟踪算法*

低信噪比条件下,单个主动或单个被动声呐节点难以实现目标跟踪。多节点声呐系统希望通过增加探测节点提高系统探测能力,但其性能提高与否取决于选择合适的数据融合算法。本文利用双节点声呐试验数据研究低信噪比条件下的目标跟踪方法,采用“结合置信度水平的表决融合”算法对多个探测节点得到的数据进行融合,既考虑了目标回波信号的信噪比特性,又考虑了目标运动的连续性特征,还考虑了各节点探测结果的决策优化,最终实现较高精度的目标跟踪。算法实现了数据级、特征级和决策级的统一、融合,通过对判决依据进行量化、分层,简化判决的复杂性。试验数据处理结果表明,该方法能较好地解决低信噪比条件下多节点声呐目标跟踪问题,目标跟踪精度较高。     

Near‐3‐MeV protons from target‐normal‐sheath‐acceleration femtosecond laser irradiating advanced targets

Advanced targets based on graphene oxide and gold thin film were irradiated at high laser intensity (10¹⁸–10¹⁹ W/cm²) with 50‐fs laser pulses and high contrast (10⁸) to investigate ion acceleration in the target‐normal‐sheath‐acceleration regime. Time‐of‐flight technique was employed with SiC semiconductor detectors and ion collectors in order to measure the ion kinetic energy and to control the properties of the generated plasma. It was found that, at the optimized laser focus position with respect to the target, maximum proton acceleration up to about 3 MeV energy and low angular divergence could be generated. The high proton energy is explained as due to the high electrical and thermal conductivity of the reduced graphene oxide structure. Dependence of the maximum proton energy on the target focal position and thickness is presented and discussed.     

含突发E层的电离层模型建立及其在测高中的应用

针对高频射线测高模型未考虑电离层突发E层(sporadic-E,Es层)的问题,从电离层物理结构特性入手,结合实测的电离层垂测数据,在多层准抛物模型的基础上研究了含Es层的电离层模型及其在目标高度测量中的应用.首先,利用反转抛物线特性模拟了含Es的电离层模型,并得到了等离子体频率与高度的关系;然后,利用该模型分析了射线的电离层传输路径与发射仰角/频率的关系以及高频射线微多径特征与目标高度的关系;最后,结合含Es的电离层模型与射线微多径特征,提出了基于分段爬山搜索的快速匹配域测高方法,该方法能大大减少搜索时间.研究结果表明:含Es的电离层模型和提出的测高方法能准确估计出目标高度,并具有较强的实时性.