基于多模差分融合的高速车辆定位算法

针对车辆高速运动下定位延时大、精度低、稳定性差的问题,发挥GPS、RSSI和INS三种定位模式的各自优势,将GPS差分校准算法与RSSI测距相结合求解横坐标,将RSSI测距与INS惯导迭代算法相结合求解纵坐标,提出了一种适合高速运动车辆的多模差分融合精确定位算法。以四车道的高速路为场景进行了仿真实验,当车速为70km/h时,跟踪误差<1m,定位延时<0.2s。实验结果表明,该算法可实现车身级和车道级定位,在精确交通诱导、车辆防撞等智能交通领域具有应用前景和推广价值。     

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。     

Biembedding Steiner Triple Systems in Surfaces Using the Bose Construction

A uniform framework is presented for biembedding Steiner triple systems obtained from the Bose construction using a cyclic group of odd order, in both orientable and nonorientable surfaces. Within this framework, in the nonorientable case, a formula is given for the number of isomorphism classes and the particular biembedding of Ducrocq and Sterboul (preprint 18pp., 1978) is identified. In the orientable case, it is shown that the biembedding of Grannell et al. (J Combin Des 6 ( 7 ), 325–336) is, up to isomorphism, the unique biembedding of its type. Automorphism groups of the biembeddings are also given.     

Classification of Graeco‐Latin Cubes

A q‐ary code of length n, size M, and minimum distance d is called an code. An code with is said to be maximum distance separable (MDS). Here one‐error‐correcting () MDS codes are classified for small alphabets. In particular, it is shown that there are unique (5, 5³, 3)₅ and (5, 7³, 3)₇ codes and equivalence classes of (5, 8³, 3)₈ codes. The codes are equivalent to certain pairs of mutually orthogonal Latin cubes of order q, called Graeco‐Latin cubes.     

Identification of Pharmacophoric Fragments of DYRK1A Inhibitors Using Machine Learning Classification Models

Dual-specific tyrosine phosphorylation regulated kinase 1 (DYRK1A) has been regarded as a potential therapeutic target of neurodegenerative diseases, and considerable progress has been made in the discovery of DYRK1A inhibitors. Identification of pharmacophoric fragments provides valuable information for structure- and fragment-based design of potent and selective DYRK1A inhibitors. In this study, seven machine learning methods along with five molecular fingerprints were employed to develop qualitative classification models of DYRK1A inhibitors, which were evaluated by cross-validation, test set, and external validation set with four performance indicators of predictive classification accuracy (CA), the area under receiver operating characteristic (AUC), Matthews correlation coefficient (MCC), and balanced accuracy (BA). The PubChem fingerprint-support vector machine model (CA = 0.909, AUC = 0.933, MCC = 0.717, BA = 0.855) and PubChem fingerprint along with the artificial neural model (CA = 0.862, AUC = 0.911, MCC = 0.705, BA = 0.870) were considered as the optimal modes for training set and test set, respectively. A hybrid data balancing method SMOTETL, a combination of synthetic minority over-sampling technique (SMOTE) and Tomek link (TL) algorithms, was applied to explore the impact of balanced learning on the performance of models. Based on the frequency analysis and information gain, pharmacophoric fragments related to DYRK1A inhibition were also identified. All the results will provide theoretical supports and clues for the screening and design of novel DYRK1A inhibitors.     

Enterocins: Classification,Synthesis, Antibacterial Mechanisms and Food Applications

Enterococci, a type of lactic acid bacteria, are widely distributed in various environments and are part of the normal flora in the intestinal tract of humans and animals. Although enterococci have gradually evolved pathogenic strains causing nosocomial infections in recent years, the non-pathogenic strains have still been widely used as probiotics and feed additives. Enterococcus can produce enterocin, which are bacteriocins considered as ribosomal peptides that kill or inhibit the growth of other microorganisms. This paper reviews the classification, synthesis, antibacterial mechanisms and applications of enterocins, and discusses the prospects for future research.     

聚光光伏系统散热研究进展

对聚光光伏系统的发展过程进行了回顾.根据电池和聚光系统分类介绍了相匹配的散热冷却方式.电池组合方式简单、聚光倍数低的系统,使用传统风冷、水冷的被动散热即可满足要求;电池组合方式复杂、聚光倍数高的系统,常利用液浸冷却、射流冲击、通道冷却或多方法组合的新型冷却方式.指出在控制成本的前提下,选择系统可靠性高、结构简单的换热方...     

排序支持的交互数据分类算法及其应用

交互分类是解决数据复杂分类问题的主要手段之一。在现有的大多交互分类系统中,用户能准确识别数据类别,但在有些分类场景中,类别之间的顺序关系更容易被识别,为此,提出一种排序支持的交互数据分类算法。为提升交互分类精度,引入数据的顺序信息,为降低标记难度,提出候选样本推荐策略。另外,提出一种评估分类算法性能的可视化方法,用包含基本车况、交通违法记录、交通事故记录等信息的车辆数据集进行实验验证,将相关车辆分为高危车辆、中危车辆、低危车辆3类,算法的分类结果模型一致度达近98%,验证了方法的有效性。     

基于多尺度学习与深度卷积神经网络的遥感图像土地利用分类

土地利用信息是国土资源管理的基础和重要依据,随着高分辨率遥感图像数据的日益增多,迫切需要快速准确的土地利用分类方法。目前应用较广的面向对象的分类方法对空间特征的利用尚不够充分,在特征选择上存在一定的局限性。为此,提出一种基于多尺度学习与深度卷积神经网络（deep convolutional neural network,DCNN）的多尺度神经网络（multi-scale neural network,MSNet）模型,基于残差网络构建了100层编码网络,通过并行输入实现输入图像的多尺度学习,利用膨胀卷积实现特征图像的多尺度学习,设计了一种端到端的分类网络。以浙江省0.5 m分辨率的光学航空遥感图像为数据源进行了实验,总体分类精度达91.97%,并将其与传统全卷积网络（fully convolutional networks,FCN）方法和基于支持向量机（support vector machine,SVM）的面向对象方法进行了对比,结果表明,本文所提方法分类精度更高,分类结果整体性更强。