Improved non-maximum suppression for object detection using harmony search algorithm

Non-maximum suppression (NMS) plays a key role in many modern object detectors. It is responsible to remove detection boxes that cover the same object. NMS greedily selects the detection box with maximum score; other detection boxes are suppressed when the degree of overlap between these detection boxes and the selected box exceeds a predefined threshold. Such a strategy easily retain some false positives, and it limits the ability of NMS to perceive nearby objects in cluttered scenes. This paper proposes an effective method combining harmony search algorithm and NMS to alleviate this problem. This method regards the task of NMS as a combination optimization problem. It seeks final detection boxes under the guidance of an objective function. NMS is applied to each harmony to remove imprecise detection boxes, and the remaining boxes are used to calculate the fitness value. The remaining detection boxes in a harmony with highest fitness value are chosen as the final detection results. The standard Pattern Analysis, Statistical Modeling and Computational Learning Visual Object Classes dataset and the Microsoft Common Objects in Context dataset are used in all of the experiments. The proposed method is applied to two popular detection networks, namely Faster Region-based Convolutional Neural Networks and Region-based Fully Convolutional Networks. The experimental results show that the proposed method improves the average precision of these two detection networks. Moreover, the location performance and average recall of these two detectors are also improved.     

一种结合区域检测和语义分割的SLAM技术

提出一种结合区域检测和语义分割的即时定位和建图（SLAM）技术，通过引入高精度图像描述子SIFT来实现前端视觉里程计（VO）过程中帧间像素匹配的精度。为了降低引入操作带来的计算复杂度，设计一个实时区域检测算法，在相邻帧间检测大致相似的ROI(Region of Interest)关键区域，使得SIFT描述子的提取和匹配只在ROI区域内完成，其余区域仍旧采用精度略低、效率更高的ORB算子。同时，为了提高后端BA(Bundle Adjustment)的精度，减少累积误差，结合语义图，在原有的基本投影误差函数上添加一个语义误差。该语义图采用实时语义分割算法完成，同时只针对ROI区域进行分割。通过与原SLAM方案对比实验，表明本文提出的方法，在提高一定精度的同时，仍能满足SLAM实时定位和建图的要求。最后，在电力作业场景下验证了该方案的效果。     

基于词法特征的恶意域名快速检测算法

针对互联网中恶意域名攻击事件频发，现有域名检测方法实时性不强的问题，提出一种基于词法特征的恶意域名快速检测算法。该算法根据恶意域名的特点，首先将所有待测域名按照长度进行正则化处理后赋予权值；然后利用聚类算法将待测域名划分成多个小组，并利用改进的堆排序算法按照组内权值总和计算各域名小组优先级，根据优先级降序依次计算各域名小组中每一域名与黑名单上域名之间的编辑距离；最后依据编辑距离值快速判定恶意域名。算法运行结果表明，基于词法特征的恶意域名快速检测算法与单一使用域名语义和单一使用域名词法的恶意域名检测算法相比，准确率分别提高1.7%与2.5%，检测速率分别提高13.9%与6.8%，具有更高的准确率和实时性。     

Homogeneous and Universal Detection of Various Targets with a Dual-Step Transduced Toehold Switch Sensor

Toehold switch sensors represent a class of new advances that allow specific targets to trigger in situ expression of a protein reporter. Although they offer unique advantages of a label-free nature and high portability, they generally require repeated sequence design, high expenditure, and laborious optimization of toehold switch sequences according to different targets. To simplify the sensing process further and to improve its practicability, we innovatively introduce a dual-step pre-transduction upon traditional toehold switch sensor. Through two successive toehold-mediated strand-displacement reactions that are initiated, respectively, by a linear and an associative trigger, different DNAs, RNAs, or ligands of functional nucleic acids can be generally transduced into the input of one high-performance toehold switch sensor. This advance greatly increases the target range. Furthermore, the whole process is signal-on, homogeneous, and free of any requirements for complicated operations such as probe labeling, separation, and reconstruction of the toehold switch, being promising and practical even in portable or point-of-care assays.     

高预应力高强度锚索支护在深井高应力巷道的应用

以红四煤矿深井高应力巷道支护为例,在研究地应力状态、围岩强度、岩石成分的前提下,通过数值模拟方式确定了高预应力高强度高刚度锚杆锚索联合支护方案。通过对试验巷道变形移近量监测发现,两帮移近变形呈稳定状态的时间提前,巷道顶底板变形状况有所改善。     

频率域电磁法在采空区探测中的应用

隐伏采空区的探测已成为近年来研究课题,采用无人机搭载发射装置,高空平面布置探测线和测点,克服在陡峭的山坡难以布置探测工作,消除地形地貌及其它条件对探测工作带的误差。对已知空区的探测和分析,验证了该方法的有效性和精确程度,为后续其它探测提供了频率域区间和参数校正。研究表明:此方法具有适用性强、工作效率高和精确度高优点,是一种可用于隐伏空区的探测方法,分析过程中可以辨别采空区变形情况。     

5通道铁磁探测仪的研发与锅炉管堵塞检测

描述了5通道铁磁探测仪的研发和应用，该仪器能在管外发现发电锅炉过热器管和再热器管内是否有铁锈堆积，并评估堵塞程度，目的是通过检测防止爆管事故。采用磁学探测方法，在管外表面的1条圆周线上布置5个测量点，进行了有限元数值模拟和标定曲线的计算。对计算结果进行了实验验证，与计算值相比较，测量值的标准误差（相对误差）为5.2%。现场检测得到的堵塞面积与管内孔的面积之比，与X射线摄影测得的结果比较，确定铁磁探测仪检测的标准误差（绝对误差）为3.9%。通过为5家热电厂进行过热器管和再热器管堵塞检测，发现超标堵塞30多处，解剖验证无一误报，所以用该仪器进行堵塞检测能够防止上述锅炉管爆管事故。     

石墨烯复合功能材料及其在重金属离子检测的电化学研究与应用

石墨烯(GR)是典型的单原子碳纳米材料,具有独特的二维共辄平面结构,其高活性的比表面积和突出的导电性能,在电催化和敏感材料制备领域已得到广泛的应用。氧化石墨烯(GO)作为GR的前驱体,存在大量的含氧官能团,具有良好的水溶分散性。大量GO含氧官能团的介入会破坏其K-7T共辘结构,导致其电学性能变差。GO通过化学、水热合成或直接电化学还原方法可有效修复其共辄平面结构,得到导电性良好的还原氧化石墨烯(rGO),即GR.单组分的GR材料在实际应用中仍存在某些局限性,如电学活性相对较弱,与其它材料加工复合性能较差等。将GR、G O材料与其它功能材料进行复合,可进一步改善复合物的物理或化学性能,如分散性、加工修饰和电催化活性等。综述了石墨烯材料与金属及其氧化物纳米粒子、聚合物、掺杂原子、导电离子液体、碳纳米材料等功能材料复合后,能形成可调控的微结构,具有改性的化学性质和协同发挥的电学效应,表现出显著的电子传递能力及其功能性作用。论述了GR功能化修饰的复合材料作为敏感界面,构筑基于重金属离子检测的电化学生物传感器,可以实现对Pb2+,Hg2+,C『+等多种重金属离子的同时或分别检出,提出了GR复合制备材料的纳米结构特征、功能修饰作用对于提高传感器的电催化活性和选择性性能等方面的应用,并对该研究领域进行了总结与展望。     

Gait event detection algorithm based on smart insoles

Gait analysis is an effective clinical tool across a wide range of applications. Recently, inertial measurement units have been extensively utilized for gait analysis. Effective gait analyses require good estimates of heel‐strike and toe‐off events. Previous studies have focused on the effective device position and type of triaxis direction to detect gait events. This study proposes an effective heel‐strike and toe‐off detection algorithm using a smart insole with inertial measurement units. This method detects heel‐strike and toe‐off events through a time‐frequency analysis by limiting the range. To assess its performance, gait data for seven healthy male subjects during walking and running were acquired. The proposed heel‐strike and toe‐off detection algorithm yielded the largest error of 0.03 seconds for running toe‐off events, and an average of 0–0.01 seconds for other gait tests. Novel gait analyses could be conducted without suffering from space limitations because gait parameters such as the cadence, stance phase time, swing phase time, single‐support time, and double‐support time can all be estimated using the proposed heel‐strike and toe‐off detection algorithm.