球面全景影像自动测量路灯坐标的方法

目的 针对城市实施的路灯杆编码项目,即获取路灯坐标并依次编号,传统的测量方法耗费大量人力物力,且作业周期长,虽然激光测量精度高,但成本也高。为此本文提出一种将基于深度学习的目标检测与全景量测结合自动获取路灯坐标的方法。方法 通过Faster R-CNN（faster region convolutional neural network）训练检测模型,对全景图像中的路灯底座进行检测,同时输出检测框坐标,并与HOG（histogram of oriented gradient）特征结合SVM（support vector machine）的检测结果进行效果对比。再将检测框的对角线交点作为路灯脚点,采用核线匹配的方式找到两幅图像中一个或多个路灯相对应的同名像点并进行前方交会得到路灯的实际空间坐标,为路灯编码做好前期工作。结果 采用上述两种方法分别对100幅全景影像的路灯进行检测,结果显示Faster R-CNN具有明显优势,采用Faster R-CNN与全景量测结合自动获取路灯坐标。由于路灯底部到两成像中心的距离大小及3点构成的交会角大小对坐标量测精度具有较大影响,本文分别对距离约为7 m、11 m、18 m的点在交会角为0°180°的范围进行量测结果对比。经验证,交会角在30°150°时,距离越近,对量测精度的影响越小。基于上述规律,在自动量测的120个路灯坐标中筛选出交会角大于30°且小于150°,距离小于20 m的102个点进行精度验证,其空间坐标量测结果的误差最大不超过0.6 m,中误差小于0.3 m,满足路灯编码项目中路灯坐标精度在1 m以内的要求。结论 本文提出了一种自动获取路灯坐标的方法,将基于深度学习的目标检测应用于全景量测中,避免手动选取全景图像的同名像点进行双像量测,节省大量人力物力,具有一定的实践意义。本文方法适用于城市车流量较少的路段或时段,以免车辆遮挡造成过多干扰,对于路灯遮挡严重的街道全景,本文方法存在一定的局限性。     

基于无人机智能视觉的输电线路全息全景重建

针对当前重建方法不能精确展现输电线路实际情况的问题,提出基于无人机智能视觉的输电线路全息全景重建方法。通过分析输电线路全息全景重建原理,采用无人机智能视觉技术对输电线路进行拍照,对输电线路全息全景图像进行采集和特征提取等预处理,跟踪图像特征,引入立体匹配算法,实现输电线路全息全景的重构。实验结果表明,所提方法输电线路重建精度高,视觉效果与实际相符,更具实用性及可行性。     

虚拟校园漫游系统面向对象程序设计的实现

全景虚拟漫游是一种基于图像绘制的虚拟现实技术,采用实景照片可以给浏览者提供一个交互性强的、身临其境的体验。介绍了现有的基于图像绘制的虚拟现实技术,讨论了使用面向对象程序设计语言设计漫游系统的可能性,最后给出了使用面向对象程序设计语言实现的太原工业学院校园漫游系统实例     

基于DaVinci~(TM)的360°全景泊车影像系统的设计与实现

全景泊车影像系统通过安装在车身前后左右的4个超广角摄像头实时采集车辆四周的影像,经过图像处理器矫正变换和全景拼接后,形成一幅车辆四周的360°全景俯视图,并实时传送到中控台的显示设备上,驾驶员坐在车中即可看到车辆所处的位置以及车辆周围的障碍物,有效减少刮蹭、碰撞、陷落等事故的发生。系统采用TI公司的低成本高性能达芬奇DSP处理器TMS320DM6437和四路视频解码器TVP5158构成系统的硬件核心,以鱼眼矫正和鸟瞰变换算法以及全景拼接技术构成系统的软件核心,实现了具有高分辨率的画质和流畅处理速度的360°全景泊车影像系统。     

The development of a panorama manifestation virtual reality system for navigation and a usability comparison with a desktop system

The virtual reality (VR) system has become popular in the last two decades and is being applied increasingly to navigation studies. This study developed a panorama manifestation (PM) VR system, with six monitors forming a circular display and an interactive chair equipped with rotation sensors and operating knobs. The advantages of the PM system include a large circular display, a body-centred design, body engagement and a low set-up cost. Based on navigation experimental tasks, this system's usability was compared with that of a typical desktop (DT) system. The results showed that participants using the PM system had a significantly higher success rate and required less completion time than participants using the DT system, indicating that the PM system outperforms the DT system in terms of effectiveness and efficiency in some navigation tasks, and also suggesting that the PM system may require lower spatial cognition workload in the navigation tasks. However, the participants’ subjective evaluations of task difficulty (TD) failed to reach significance although the PM group did report a lower mean value for TD. Additionally, in the PM system, the participants’ spatial ability was more predictive to their navigation performance than that in the DT system, indicating that the PM system offered greater usability for some spatial experiments and could provide more support for participants’ navigation tasks.     

Penalized cluster analysis with applications to family data

The goal of cluster analysis is to assign observations into clusters so that observations in the same cluster are similar in some sense. Many clustering methods have been developed in the statistical literature, but these methods are inappropriate for clustering family data, which possess intrinsic familial structure. To incorporate the familial structure, we propose a form of penalized cluster analysis with a tuning parameter controlling the tradeoff between the observation dissimilarity and the familial structure. The tuning parameter is selected based on the concept of clustering stability. The effectiveness of the method is illustrated via simulations and an application to a family study of asthma.     

Circular intra prediction for 360 degree video coding

Different from traditional 2D video, the contents of 360 degree video are deformed due to the projection from 3D sphere to 2D plane. As a result, the traditional Angular Intra Prediction (AIP) with a linear pattern may not be always efficient. To further improve the coding performance of 360 degree video, a novel intra prediction method is presented in this paper, i.e., Circular Intra Prediction (CIP), which takes consideration of the spherical characteristics of 360 degree video. In specific, the proposed CIP is performed in a circular pattern, where the center of circle is located around the to-be-predicted block, and different centers of circle are able to produce different CIP modes. The distance between center of this circle and center of the to-be-predicted block is adaptively determined according to the degree of projection deformation, where stronger projection deformation needs shorter distance, and vice versa. As the increase of the distance, the CIP is more and more close to the traditional AIP. In addition, one additional binary flag is utilized to achieve better coding performance from the competition between AIP and CIP with the rate-distortion optimization. The proposed algorithm is implemented on the platform of Versatile video coding Test Model (VTM) 5.0 + 360Lib 9.1. Extensive experiments show that the proposed method can achieve bit rate reduction on this platform for 360 degree video coding.     

攀钢360m~2烧结机工艺技术的优化

为实现攀钢烧结机的技术升级,2009年6月建成投产了1台360m2的大型烧结机。达产后由于生产不稳定、设备故障率高、指标落后等问题影响了高产稳产,为此开展了工艺技术优化研究。通过优化工艺流程,改进原燃料质量、烧结矿成分优化以及提高工艺参数控制水平等措施,利用系数达到1.414t/（m2.h）,转鼓指数提高到73.27%,固体燃耗45.31kg/t,作业率96.11%。     

四方V360变频器在吹膜机收卷上的应用

本文介绍了四方V360矢量型变频器在吹膜机收卷上的成功应用。