Looking ahead: Joint small group detection and tracking in crowd scenes

Small group detection and tracking in crowd scenes are basis for high level crowd analysis tasks. However, it suffers from the ambiguities in generating proper groups and in handling dynamic changes of group configurations. In this paper, we propose a novel delay decision-making based method for addressing the above problems, motivated by the idea that these ambiguities can be solved using rich temporal context. Specifically, given individual detections, small group hypotheses are generated. Then candidate group hypotheses across consecutive frames and their potential associations are built in a tree. By seeking for the best non-conflicting subset from the hypothesis tree, small groups are determined and simultaneously their trajectories are got. So this framework is called joint detection and tracking. This joint framework reduces the ambiguities in small group decision and tracking by looking ahead for several frames. However, it results in the unmanageable solution space because the number of track hypotheses grows exponentially over time. To solve this problem, effective pruning strategies are developed, which can keep the solution space manageable and also improve the credibility of small groups. Experiments on public datasets demonstrate the effectiveness of our method. The method achieves the state-of-the-art performance even in noisy crowd scenes.     

彩色汽车牌照的定位方法

基于颜色和纹理综合特征的车牌定位方法,通过色彩空间变换、色彩分割,牌照区域定位及去除伪车牌区域等步骤进行车牌特征分析.其在HSV空间中进行色彩分割,运用数学形态学处理,采用闭运算,形成连通域.通过连通域分析车牌区域,再根据其几何和空间频率特征,设计相关评判标准.经实验验证,符合该标准的即为车牌区域.     

时变滑模变结构与智能模糊结合的非线性系统控制

时变滑模变结构与智能模糊结合的非线性系统控制,先采用时变滑模变结构控制,通过设计动态滑模面,系统的任意初始状态一开始就处于系统的滑模面上,使系统对内部参数变化和外部干扰均具有全局鲁棒性.然后设计智能模糊控制器,使整个系统的鲁棒性在线调整,不仅增强了全局抗干扰能力,而且有效消除系统的抖振现象.     

Novel wide band-gap polymer utilizing fused hetero-aromatic unit for efficient polymer solar cells and field-effect transistors

Novel wide band-gap polymer of PBTFT containing dibenzosexithiophene-alt-bithiophene backbone was designed and synthesized via the Stille cross-coupling reaction. This polymer exhibited good thermal stability, well coplanar backbone and a broad absorption band from 350 nm to 610 nm with a wide optical band-gap of 2.02 eV. The polymer solar cells (PSCs) based on the PBTFT:PC₇₁BM active layer showed the power conversion efficiency of 3.0% with an open circuit voltage of 0.70 V, a short-circuit current of 7.94 mA cm⁻² and a fill factor of 53.98% under the illumination of AM.1.5, 100 mW cm⁻². Holes mobility up to 0.028 cm² V⁻¹ s⁻¹ with an on-off ratio of 1.0 × 10⁶ was obtained in the PBTFT-based organic field-effect transistors (OFETs). Our work indicates that the dibenzosexithiophene-alt-bithiophene based copolymer can be efficiently applied in PSCs and OFETs.     

Spin filtering effect and magnetoresistance in zigzag 6, 6, 12-graphyne nanoribbon system

The electronic transport properties of zigzag 6, 6, 12-graphyne nanoribbons (6, 6, 12-Z GYNRs) are investigated by first-principles calculations. The results show that the two bands around the Fermi level of 4-6, 6, 12-ZGYNR are not flat but cambered, which is different from that of zigzag graphene and α-graphyne nanoribbons, and is also different from that of 5-6, 6, 12-ZGYNR. In non-magnetic states, the current across the 5-6, 6, 12-ZGYNR system is almost forbidden even at rather large bias voltages, while in 4-6, 6, 12-ZGYNR system, the current increases linearly with the increase of bias voltage. In spin-polarized state, fine spin filtering effect is gained in anti-parallel configuration of 5-6, 6, 12-ZGYNR system. Moreover, it is found that magnetoresistance exists in the 5-6, 6, 12-ZGYNR system, but not in the 4-6, 6, 12-ZGYNR system.     

基于分段加权的反向稀疏跟踪算法研究

为提高稀疏表示跟踪模型性能，提出一种分段加权的反向稀疏跟踪算法，将跟踪问题转化为在贝叶斯框架下寻找概率最高的候选对象问题，构造不同的分段权重函数来分别度量候选目标与正负模板的判别特征系数。通过池化来降低跟踪结果的不确定性干扰，选择正负模板加权系数差值最大的候选表示作为跟踪结果。实验表明，在光照变化、遮挡、快速运动、运动模糊情况下，所提出的算法可以确保跟踪结果的准确性和鲁棒性。     

Synthesis and photovoltaic properties of D-A copolymers based on alkylthio-thiophene or alkylthio-selenophene-BDT donor unit and DPP acceptor unit

Two new 2D-conjugated D-A copolymers, PBDTT-S-DPP and PBDTSe-S-DPP, based on benzodithiophene (BDT) donor unit with alkylthio-thiophene or alkylthio-selenophene conjugated side chains and 2,5-bis(2-butyloctyl)-3,6-di(thiophen-2-yl)pyrrolo[3,4-c]pyrrole-1,4(2H,5H)-dione) (DPP) acceptor unit, were synthesized for the application as donor materials in polymer solar cells (PSCs). The two polymers were characterized by absorption spectroscopy, cyclic voltammetry, thermogravimetric analysis, theoretical calculation with density functional theory, X-ray diffraction and photovoltaic measurements. The results show that the alkylthio-thiophene/selenophene side groups on BDT unit and intramolecular hydrogen bonding interaction in DPP acceptor unit play important roles in affecting the absorption, HOMO energy levels, molecular planarity and the crystallinity of the polymers. The PSCs based on PBDTT-S-DPP or PBDTSe-S-DPP as donor and PC₇₁BM as acceptor demonstrate power conversion efficiency (PCE) of 5.62% and 5.01%, with relatively higher V_oc of 0.79 V and 0.76 V, respectively.     

Influence of propyl alcohol addition during processing of molybdenum trioxide powders: Observation of novel dual-wavelength excitation photochromism

The novel dual-wavelength excitation photochromic molybdenum trioxide (MoO₃) powders are fabricated successfully via the hydrothermal method by the addition of propyl alcohol. The samples were fully characterized by means of X-ray diffraction (XRD), scanning electron microscopy (SEM), ultraviolet–visible diffuse reflectance spectroscopy (UV–vis), and colorimetry. The characteristic spectrum of XRD demonstrates that the as-obtained products are highly pure. The results show that the flowerlike MoO₃ sample with the inducer of propyl alcohol has better photochromic properties, which was analyzed by colorimetry. It is shown that both of the samples exhibit a strong adsorption band between 250 and 400 nm in the UV range. Moreover, the UV–vis spectra shows that MoO₃ synthesized with added propyl alcohol absorbs light not only between 250 and 400 nm but also displays an additional band between 500 and 800 nm, demonstrating novel dual-wavelength excitation photochromic properties.     

基于视觉感知的智能扫地机器人的垃圾检测与分类

目的 为了提高扫地机器人的自主性和智能化程度，为扫地机器人配备视觉传感器，使其获得视觉感知能力，通过研究有效的垃圾检测分类模型与算法，实现对垃圾的定位与识别，引导扫地机器人对垃圾进行自动识别与按类处理，提高工作的目的性和效率，避免盲动和减少能耗。方法 选择检测速度较快的YOLOv2作为主网络模型，结合密集连接卷积网络，嵌入深层密集模块，对YOLOv2进行改进，提出一种YOLOv2-dense网络，该网络可以充分利用图像的高分辨率特征，实现图像浅层和深层特征的复用与融合。结果 测试结果表明，智能扫地机器人使用本文方法可以有效识别不同形态的常见垃圾类别，在真实场景中，测试识别准确率为84.98%，目标检测速度达到26帧/s。结论 实验结果表明，本文构建的YOLOv2-dense网络模型具有实时检测的速度，并且在处理具有不同背景、光照、视角与分辨率的图片时，表现出较强的适应和识别性能。在机器人移动过程中，可以保证以较高的准确率识别出垃圾的种类，整体性能优于原YOLOv2模型。