融合明度特征的V-SAD立体匹配算法

针对SAD算法难以对图像中纯色与弱纹理部分进行准确匹配的问题,提出了将HSV空间明度特征与SAD算法相融合的立体匹配算法,称为V-SAD算法.首先将图像从RGB空间转化至HSV空间,并根据H、S、V值将像素点按照颜色分为10类,同时得到明度特征图.然后结合SAD算法需要的灰度特征图计算匹配代价,得到初步的视差图.接着,根据HSV空间的颜色信息对图像进行分割,结合数学形态法求解分割区块中的独立连通域.再利用边缘生长法对每一个连通域的视差进行恢复.最后,使用左右一致性检测方法对视差图进行优化.实验结果表明,利用图像的明度信息衡量纯色与弱纹理区域的匹配点的相似度是有效的,该V-SAD算法有效改善了SAD算法在弱纹理和纯色区域的匹配效果,平均误匹配率下降了13.02％.     

Target localization based on cross-view matching between UAV and satellite

Matching remote sensing images taken by an unmanned aerial vehicle (UAV) with satellite remote sensing images with geolocation information. Thus, the specific geographic location of the target object captured by the UAV is determined. Its main challenge is the considerable differences in the visual content of remote sensing images acquired by satellites and UAVs, such as dramatic changes in viewpoint, unknown orientations, etc. Much of the previous work has focused on image matching of homologous data. To overcome the difficulties caused by the difference between these two data modes and maintain robustness in visual positioning, a quality-aware template matching method based on scale-adaptive deep convolutional features is proposed by deeply mining their common features. The template size feature map and the reference image feature map are first obtained. The two feature maps obtained are used to measure the similarity. Finally, a heat map representing the probability of matching is generated to determine the best match in the reference image. The method is applied to the latest UAV-based geolocation dataset (University-1652 dataset) and the real-scene campus data we collected with UAVs. The experimental results demonstrate the effectiveness and superiority of the method.     

地图辅助行人航迹推算技术的室内定位方法

针对目前室内定位技术中位置漂移、定位结果偏差较大的问题,提出了一种融合行人航迹推算(PDR)、地标匹配修正和地图辅助的室内行人定位方法.该方法利用基于惯性传感器的PDR技术计算行人位置信息,利用行走过程中的传感器读数特征识别室内特定地标点,与地标库数据进行匹配后,修正PDR轨迹产生的累积误差.室内地图辅助主要是通过判断行人是否位于走廊等区域,限制轨迹穿墙,约束PDR定位轨迹.实验结果表明,融合定位算法得到的轨迹优于纯惯性递推算法得到的轨迹,更加接近真实的行走轨迹,定位精度提高了51.2％,平均定位误差降至1.8m,满足室内定位需求.     

基于注意力机制和特征匹配的空对地多目标检测与跟踪算法

多目标跟踪算法是实现无人机自主导航的关键技术，为解决现有方法存在的小目标检测能力弱、计算能耗大、鲁棒性差等问题，提出一种基于注意力机制和特征匹配的多目标空对地跟踪算法，以实现航拍视角下对目标的精准高效跟踪。首先，引入通道可分离卷积，实现目标检测模型的轻量化；其次，构造融合空间注意力机制的小目标检测分支，提高对小微目标的检测精度，最后，优化目标跟踪算法的外观重识别网络，提高多目标跟踪效率。使用Visdrone2019-MOT数据集对所提算法进行验证，实验结果表明，所提算法的MOTA值提高了0.６%，FPS值为21.31帧/s，在模型大小和跟踪精度上实现了较好的平衡。     

共沉淀法制备花状NiCo2O4及其吸波性能的研究

通过共沉淀法制备花状NiCo₂O₄吸波材料，利用XRD、FTIR、BET、XPS和SEM等方式表征了样品成分及形貌特征，将不同煅烧温度（300、400、500、600、700 ℃）的产物利用矢量网络分析仪通过同轴测试法模拟了不同厚度下的吸波性能。结果表明：400 ℃煅烧样品在反射损耗值、吸收带宽以及样品厚度都表现出较好的性能，当厚度为1.5 mm时，在14.32 GHz处吸波材料达到最大的吸收损耗值为-43.71 dB，有效吸收带宽为4.48 GHz （12.08~16.56 GHz），可以预见镍钴基吸波材料具有很大的潜力。     

A geographical and operational deep graph convolutional approach for flight delay prediction

Flight delay prediction has attracted great interest in civil aviation community due to its significant role in airline planning, flight scheduling, airport operation, and passenger service. Flight delay is affected by numerous factors and irregularly propagates in air transportation networks owing to flight connectivity, which brings critical challenges to accurate flight delay prediction. In recent years, Graph Convolutional Networks (GCNs) have become popular in flight delay prediction due to the advantage in extracting complicated relationships. However, most of the existing GCN-based methods have failed to effectively capture the spatial–temporal information in flight delay prediction. In this paper, a Geographical and Operational Graph Convolutional Network (GOGCN) is proposed for multi-airport flight delay prediction. The GOGCN is a GCN-based spatial–temporal model that improves node feature representation ability with geographical and operational spatial–temporal interactions in a graph. Specifically, an operational aggregator is designed to extract global operational information based on the graph structure, while a geographical aggregator is developed to capture the similar nature among spatially close airports. Extensive experiments on a real-world dataset demonstrate that the proposed approach outperforms the state-of-the-art methods with a satisfying accuracy improvement.     

液气式针栓喷注单元喷雾场结构试验研究

径向孔型针栓喷注器相对于径向缝型针栓喷注器具有更复杂的喷雾场。为了研究径向孔型针栓喷注器的喷雾场结构，将径向孔型针栓喷注器简化为单个气体射流与液膜碰撞的针栓喷注单元，采用了背景光成像系统结合激光相位多普勒技术（PDA），以水和空气为模拟介质，对液气式针栓喷注单元的喷雾场进行了试验研究。试验结果表明，液气式针栓喷注单元喷雾的三维结构呈现“喇叭”状。根据喷雾的形成过程及液滴的分布，液气式针栓喷注单元喷雾可以划分为4个区域：碰撞区、液滴区、液雾区及液丝区。液气式针栓喷注单元喷雾的分布范围可由内边界角、外边界角、中线角及散布角表示，均随局部动量比的增大而增大。液滴区的粒子主要由碰撞过程产生，SMD较大；液雾区的粒子经碰撞过程产生后，在气动力作用下进一步雾化，SMD小。由于液雾区的速度和粒径同时受到气动力作用的影响，粒径分布与速度分布在空间上呈现负相关趋势。