YOLSO: You Only Look Small Object

Small object detection is challenging and far from satisfactory. Most general object detectors suffer from two critical issues with small objects: (1) Feature extractor based on classification network cannot express the characteristics of small objects reasonably due to insufficient appearance information of targets and a large amount of background interference around them. (2) The detector requires a much higher location accuracy for small objects than for general objects. This paper proposes an effective and efficient small object detector YOLSO to address the above problems. For feature representation, we analyze the drawbacks in previous backbones and present a Half-Space Shortcut(HSSC) module to build a background-aware backbone. Furthermore, a coarse-to-fine Feature Pyramid Enhancement(FPE) module is introduced for layer-wise aggregation at a granular level to enhance the semantic discriminability. For loss function, we propose an exponential L1 loss to promote the convergence of regression, and a focal IOU loss to focus on prime samples with high classification confidence and high IOU. Both of them significantly improves the location accuracy of small objects. The proposed YOLSO sets state-of-the-art results on two typical small object datasets, MOCOD and VeDAI, at a speed of over 200 FPS. In the meantime, it also outperforms the baseline YOLOv3 by a wide margin on the common COCO dataset.     

Influence of temperature on acoustic emission source location accuracy in underground structure

An acoustic emission (AE) experiment was carried out to explore the AE location accuracy influenced by temperature. A hollow hemispherical specimen was used to simulate common underground structures. In the process of heating with the flame, the pulse signal of constant frequency was stimulated as an AE source. Then AE signals received by each sensor were collected and used for comparing localization accuracy at different temperatures. Results show that location errors of AE keep the same phenomenon in the early and middle heating stages. In the later stage of heating, location errors of AE increase sharply due to the appearance of cracks. This provides some beneficial suggestions on decreasing location errors of structural cracks caused by temperature and improves the ability of underground structure disaster prevention and control.     

Charge source and the charging mechanism of the contact electrification of polymer powder

The charge sources, as well as the charging mechanism of the contact electrification (CE) of polymers, are still debatable. Since CE is accompanied by destruction, it is considered that “hard contacting” via ball milling can induce covalent bond scission and produce naked-activated-charge sources. Regarding “soft contacting” via nano-scale sliding, which does not induce covalent bond scission, a frontier-electron, “f-electron,” of the naked-activated-charge source is crucial to electron transfer among the naked-activated-charge sources. Here, we configure naked-activated-charge-source models, naked-activated-mechano-anion, and naked-activated-mechano-cation, which are produced by mechanical energy induced heterogeneous covalent bond scission, as well as naked-activated-mechano-radicals that are produced by homogeneous covalent bond scission. Regarding “soft contacting” among naked-activated-charge sources in a vacuum, f-electron can be transferred from a donor to an acceptor if the energy level of the donor is higher than that of the acceptor. The net amount of the normalized transferred-f-electrons is obtained by adopting settings in which the average energy level of the naked-activated-charge sources (as the donors) is higher than that of the sources employed as acceptors. Thus, the surfaces comprising the donors and acceptors will exhibit positive and negative net surface charges, respectively. We conclude that net surface charges depend on the average energy level of naked-activated-charge sources. Further, we observe that the alignment of polyethylene (PE)-polyvinyl chloride (PVC)-polytetrafluoroethylene (PTFE) to the average energy level is identical to that of the triboelectric series.     

基于声源定位的高校在线巡课系统设计与研究

随着网络建设以及信息化教学方法在高校教学过程中的应用普及,越来越多的高校使用在线巡课系统对教师的教学过程进行跟踪和管理,以便发现课堂教学中的亮点、问题和不足。文章提出的在线巡课系统,基于声源定位的技术,对传统的在线巡课系统做出了改进,解决了已有巡课系统中“只闻其声,不见其人”的问题,能够更直观地跟踪到教师的教学过程,包括教学行为以及师生互动过程,有效提升教务人员巡课效果和体验感受。     

Inverse identification of constitutive parameters from heat source fields: A local approach applied to hyperelasticity

In this paper, a new inverse identification method of constitutive parameters is developed from full kinematic and thermal field measurements. It consists in reconstructing the heat source field from two different approaches by using the heat diffusion equation. The first one requires the temperature field measurement and the value of the thermophysical parameters. The second one is based on the kinematic field measurement and the choice of a thermo-hyperelastic model that contains the parameters to be identified. The identification is carried out at the local scale, ie, at any point of the heat source field, without using the boundary conditions. In the present work, the method is applied to the challenging case of hyperelasticity from a heterogeneous test. Due to large deformations undergone by the rubber specimen tested, a motion compensation technique is developed to plot the kinematic and the thermal fields at the same points before reconstructing the heterogeneous heat source field. In the present case, the constitutive parameter of the Neo-Hookean model has been identified, and its distribution has been characterized with respect to the strain state at the surface of a cross-shaped specimen.     

综放高瓦斯矿井煤层瓦斯涌出量综合预测

为了成功预测竹林山煤矿综放高瓦斯矿井大采高工作面煤层瓦斯涌出量,以主采3号煤层为主要研究对象,针对3号煤层以往开采情况,通过布设测点测量其煤层瓦斯含量和了解相邻矿井瓦斯含量,采用分源预测法、回归法及统计法等预测方法得到了3号煤层瓦斯含量的分布规律,并绘制了3号煤层的瓦斯含量等值线图。对矿井不同生产时期的瓦斯含量进行预测,得到了生产前期、中期及后期采区的最大绝对瓦斯涌出量和最大相对瓦斯涌出量,说明了竹林山煤矿各个时期均属于高瓦斯矿井。     

Study of planar structure applying reciprocity technique combined with MGEC and analysis of discontinuity

The 3D structure electromagnetic computation presents several difficulties related to the volume mesh. In fact, the entire volume space must be taken into account even the smallest details. In this article, we propose a formulation based on the reciprocity theorem combined with the generalized equivalent circuit method to model a planar 3D structure with both coaxial and planar excitation. The major advantage of this formulation is the fact to reduce the computational volume into 2D ones in the discontinuity plane. In addition, we focused on the calculation of the discontinuity between the excitation source and the planar structure to determine the exact behavior of the electric coaxial excitation model. The obtained current density, electric field distributions, and the input impedance are presented and discussed in the following sections. An approximately good agreement of input impedance with those obtained by the simulator and measurement is shown.     

The generalized discrete ‐centroid problem

The ‐centroid problem or leader–follower problem is generalized considering different customer choice rules where a customer may use facilities belonging to different firms, if the difference in travel distance (or time) is small enough. Assuming essential goods, some particular customer choice rules are analyzed. Linear programming formulations for the generalized ‐medianoid and ‐centroid problems are presented and an exact solution approach is applied. Some computational examples are included.     

Restructuring bank networks after mergers and acquisitions: A capacitated delocation model for closing and resizing branches

During restructuring processes, due to mergers and acquisitions, banks frequently face the problem of having redundant branches competing in the same market. In this work, we introduce a new Capacitated Branch Restructuring Model which extends the available literature in delocation models. It considers both closing down and long term operations׳ costs, and addresses the problem of resizing open branches in order to maintain a constant service level. We consider, as well, the presence of competitors and allow for ceding market share whenever the restructuring costs are prohibitively expensive.We test our model in a real life scenario, obtaining a reduction of about 40% of the network size, and annual savings over 45% in operation costs from the second year on. We finally perform a sensitivity analysis on critical parameters. This analysis shows that the final design of the network depends on certain strategic decisions concerning the redundancy of the branches, as well as their proximity to the demand nodes and to the competitor׳s branches. At the same time, this design is quite robust to changes in the parameters associated with the adjustments on service capacity and with the market reaction.     

Application of differential evolution for cascaded multilevel VSI with harmonics elimination PWM switching

Harmonic elimination pulse width modulation (HEPWM) method has been widely applied to multilevel voltage source inverter (MVSI) to remove low frequency harmonics from its output voltage. However, the computation of the HEPWM switching angles for MVSI is very challenging due to several constraints, namely angle sequencing, very tight angular spacing and the numerous possibilities of angles distribution ratio. Realizing the potential of Differential Evolution (DE) to handle complex problems, this work proposes its application to solve the HEPWM problem for cascaded MVSI. Its emphasis is on improving the availability of HEPWM for higher output voltage by extending the maximum range of modulation index (M). It also removes the discontinuities in the switching angles and reduces the number of distribution ratio required to obtain the required solution. Compared to the most advanced (similar) work, i.e., 7-level MVSI with seventeen switching angles, DE covers a wider range of M; the maximum achievable M is 2.80. Furthermore, it exhibits very low second order distortion factor (DF₂): for the worst case, the value of DF₂ is 0.0014%. To verify the viability of the proposed algorithm, simulation is carried out and hardware prototype is constructed. Both results show very good agreement with the theoretical prediction.