Visual lobe shape characteristics of experienced industrial inspectors and inexperienced subjects

This article reports research on visual lobe shape differences between experienced industrial inspectors and inexperienced students and the effects of years of inspection experience on lobe‐shape characteristics. Comparison of the visual lobes of students and inspectors showed that lobe roundness of the inspectors was higher than that of the students. Also, marked improvement in lobe roundness and regularity for inspectors was evident in those inspectors who had relatively long experience in inspection work, indicating that daily inspection practice did not seem to be an effective way of improving lobe‐shape characteristics. The increased lobe roundness level did not change, however, even when the inspectors had not performed an inspection job for a long period of time. This finding suggests that the improved lobe‐roundness level could be maintained even without any follow‐up inspection practice. It seems that intensive lobe training rather than inspection practice may be required to improve lobe‐shape characteristics in a short time. No relationship was found between lobe‐shape parameters and the ranked inspection performance of the inspectors, which indicated that, for the inspectors, better search performance may be related more to their better decision making rather than to their higher lobe‐roundness level or inspection speed. © 2010 Wiley Periodicals, Inc.     

轨检小车在高速铁路轨道检测及精调中的应用

轨道线路参数直接影响列车的运行安全和旅客的舒适性,目前轨检小车在高速铁路轨道施工及后期运营维护已广泛使用。本文结合广州南方高铁轨检小车在成渝客专、成贵高铁等项目施工过程中的轨道精测精调实践,主要对轨检小车在高速铁路轨道检测作业进行归纳总结,从轨道测量技术、轨检小车检测原理、作业流程、施工注意要点等多方面提出具有指导意义的结论。     

无人机影像目标检测技术研究进展

无人机影像目标检测技术是无人机影像智能化分析中的关键技术，开展无人机影像目标检测技术的研究有着广阔的应用前景和价值。介绍了无人机影像中目标检测技术的发展历程，简述了无人机影像目标检测技术在森林防火、农业信息采集、电力线路检测、路桥损害检测以及军事侦察等5种特定领域的应用情况，总结归纳了目标检测性能评价时常用的数据集和标准，并对未来无人机影像目标检测技术的发展态势进行了展望。     

基于深度可分离卷积的地铁隧道巡检视频分析

地铁隧道安全目前主要依靠地铁轨道巡检员在轨道无车时人工巡轨检查,这种方法速度慢、工作效率低,而且巡检效果完全依赖于轨道巡检员的经验和状态。针对这一问题,提出了一种基于深度可分离卷积的地铁隧道巡检视频异常报警系统,该系统使用提出的SubwayNet卷积神经网络完成对巡检视频图像的分类。利用制作的地铁隧道巡检数据集对构建的卷积神经网络进行训练并保存模型,制作了图形用户界面并加入声音报警的功能,最后将程序文件打包为可执行文件。实验结果表明,该系统的分类准确率能够达到96%,速度能够达到52 fps,满足对视频实时、准确分析的要求。     

无人机航拍图像语义分割研究综述

随着无人机技术的快速发展,无人机在研究领域和工业应用方面受到了广泛的关注。图像和视频是无人机感知周围环境的重要途径。图像语义分割是计算机视觉领域的研究热点,在无人驾驶、智能机器人等场景中应用广泛。无人机航拍图像语义分割是在无人机航拍图像的基础上,运用语义分割技术使无人机获得场景目标智能感知能力。介绍了语义分割技术和无人机的应用发展、相关无人机航拍数据集、无人机航拍图像特点和常用语义分割评价指标。针对无人机航拍的特点介绍了相关语义分割方法,包括小目标、模型实时性和多尺度整合等方面。综述无人机语义分割相关应用,包括线检测、农业和建筑物提取等方向,并分析无人机语义分割未来发展趋势和挑战。     

输电线路巡检无人机导航的语义交互方法

针对高空巡检环境的复杂性和不确定性,提出了一种输电线路巡检无人机导航的语义交互方法.基于本体方法构建人机交互的语义拓扑体系,形成具有感知运动数据处理的导航框架;其次,以语义拓扑关系为基础,结合谓词逻辑形式的背景知识,提出结构化路径导航方法,实现人机在统一的概念层次上路径导航的语义交互;最后研发了巡检导航的语义交互原型系统,通过实验证明：作为语义交互实例,本文所提方法可以为无人机电力巡线提供位置导航控制,实体名词提取满足实际期望估计且精度控制较高,为基于数据驱动的输电线路巡检提供了基础支持.     

面向无人机电力巡检的语义实体构建及航迹控制方法

航迹的合理控制是影响无人机（UAV）智能决策重要因素。考虑UAV巡检的局部观测性和任务环境的高空复杂性，以电力巡检领域知识为背景，提出面向UAV电力巡检的语义实体构建及航迹控制方法。首先，基于电力巡检领域的实体知识构建空间拓扑网络，并生成关于位置节点的语义航迹序列网络及其语义接口；然后，根据空间拓扑结构相似性度量的结果集，提出安全许可机制和基于强化学习的航迹控制策略，实现UAV电力巡检在统一的概念内涵和位置结构上的轨迹控制。实验结果表明：作为UAV巡检的实例，所提方法得到的最优策略能获得最大化的鲁棒性能；同时，该方法通过强化学习方法使目标网络的适应度稳定收敛且实体区域覆盖率高于95%，为UAV电力巡检任务决策提供了飞行依据。     

面向斜拉桥索塔巡检的旋翼UAV避障航迹规划

针对斜拉桥索塔巡检的旋翼UAV避障航迹规划问题,提出了一种面向斜拉桥索塔巡检的旋翼UAV避障航迹规划方法。该方法以巡检过程中旋翼UAV的能量消耗为航迹优劣评价指标,利用基于信息熵理论改进后的遗传算法获取能量消耗最少航迹,并提出双圆弧避障策略对航迹上存在斜拉索障碍的局部区域进行航迹重新规划,使之能有效地避让斜拉索障碍,保障旋翼UAV的飞行安全。以咸阳渭城桥索塔的外观巡检为例进行仿真验证,仿真结果表明,所提方法规划的航迹有效地降低了巡检旋翼UAV的能量消耗,确保了巡检旋翼UAV的飞行安全,能够较好地适用于斜拉桥索塔外观巡检。     

面向无人机电力巡检的语义实体构建及航迹控制方法

航迹的合理控制是影响无人机（UAV）智能决策重要因素。考虑UAV巡检的局部观测性和任务环境的高空复杂性,以电力巡检领域知识为背景,提出面向UAV电力巡检的语义实体构建及航迹控制方法。首先,基于电力巡检领域的实体知识构建空间拓扑网络,并生成关于位置节点的语义航迹序列网络及其语义接口;然后,根据空间拓扑结构相似性度量的结果集,提出安全许可机制和基于强化学习的航迹控制策略,实现UAV电力巡检在统一的概念内涵和位置结构上的轨迹控制。实验结果表明：作为UAV巡检的实例,所提方法得到的最优策略能获得最大化的鲁棒性能;同时,该方法通过强化学习方法使目标网络的适应度稳定收敛且实体区域覆盖率高于95%,为UAV电力巡检任务决策提供了飞行依据。     

基于S7—200PLC的十吨级龙门吊自动转向控制系统

十吨级龙门吊应用于高速铁路轨道板的铺设作业中,为保证龙门吊在桥上安全行驶,桥的两侧设置了防撞墙,设计一套龙门吊自动转向系统,使龙门吊在走行过程中与防撞墙保证安全距离。研究并试验了以PLC为控制器,由超声测距传感器、LVDT位移传感器完成位置检测,并用电液推杆完成转向动作的自动转向系统。实验结果表明系统可以实现上述要求,控制策略可行,控制系统工作平稳、结构紧凑、易于调试。     

救灾无人机的优化A*航迹规划算法

针对抢险救灾中无人机派遣量及空间航迹规划最短路径相制约的问题,提出了一种优化A*的航迹算法。通过设计的蛇形割圆法对圆形巡查区域进行路径规划,通过提取感兴趣区域的方法选择较佳搜索方向,提高搜索效率,采用加权评估法优化自然威胁权重系数,重定义航迹估计函数。将提出的方法在灾情巡查和生命勘测实际问题中进行性能检测。仿真结果表明,该算法能够合理分配无人机数量且能快速规划出较优飞行轨迹,实现巡查覆盖率达88.96%。     

无人机视觉SLAM环境感知发展研究

无人机在进行搜索救援等高级任务的时候,往往需要确定自己的位置和环境信息;仿照于人类通过视觉感知环境信息,视觉SLAM是计算机视觉领域里面通过视觉传感器感知环境的信息并快速跟踪自身的位置和建立环境地图的一种前沿技术;文章首先阐述了 VSLAM的重要组成部分:前端处理(特征点法和直接法)、数据关联、后端优化算法(滤波方法和优化方法)和建图;然后总结了一些在无人机上成功应用的典型VSLAM算法,以及在VSLAM发展的30多年的时间里涌现出许多出色的方案和研究机构;接着论述了当前用于无人机VSLAM发展的几个重点问题,多无人机协同的C—SLAM、深度学习和语义分割在SLAM中的应用、以视觉惯导为代表的多传感器融合SLAM;最后,对VSLAM方法进行总结,给出了未来的发展方向,希望对后续研究提供指导和帮助.     

Differences between inexperienced and experienced safety supervisors in identifying construction hazards: Seeking insights for training the inexperienced

The hazard identification ability of frontline safety supervisors is essential to ensure site safety. As experience can benefit the identification performance, this study investigates the gaps between inexperienced and experienced safety supervisors. Thirty-five experienced safety supervisors and 35 novices were invited to identify hazards in 18 virtual construction sites created by 360-degree panoramas. Their identification results, attention allocation, and adopted scanpaths during the identification process were compared. It is found that the experienced significantly spent more fixation time, had more fixations, and gave a larger proportion of attention to hazardous areas. In contrast, the inexperienced had no idea about where might exist hazards in a scenario. They missed hazards due to ignoring the hazardous areas. Besides, it was hard for the inexperienced to recognize hazards requiring in-depth knowledge of safety regulations. They significantly identified fewer hazards except for the relatively obvious hazards: improper use of PPE and struck-by hazards. The scanpaths were more consistent among the experienced. They observed the scene sequentially, without consciously adopting any specific searching patterns from which the novices could learn. Therefore, it is suggested to train the inexperienced to be aware of hazardous areas in workplaces in addition to educating them on safety norms; and provide them chances to practice hazard identification to retain their learned knowledge. The findings reveal the gaps between inexperienced and experienced safety supervisors, providing insights for training the inexperienced and thus helping ensure the job site safety.     

基于优化SSD的低空无人机检测方法

针对浅层特征缺乏语义信息和小目标特征不显著的问题,提出了一种基于多尺度特征融合和注意力的低空无人机（unmanned aerial vehicle,UAV）检测方法。首先提出一种多尺度特征融合模块,将不同尺度的特征图进行有效融合,使浅层特征图的细节纹理信息和深层特征图的语义信息得到充分的利用,改善浅层特征语义信息不足的问题。然后在网络特征图输出处引入一种不降维局部跨信道交互策略和核大小自适应选择的通道注意力机制,以极其轻量级的方式获取跨通道的交互信息。为使先验框和有效感受野匹配,优化默认框设置方法,更好地检测小目标。使用自制无人机数据集进行验证,结果表明改进后算法平均准确率为84.07%,比原始SSD（single shot multibox detector）算法提高了7.81个百分点,检测速度达到31.3?frame/s。     

基于完全自注意力的水电枢纽缺陷识别方法

水电枢纽在长期运行过程中容易受水流侵蚀、应力变化等因素影响,导致形成裂缝、渗漏、脱落、露筋等缺陷,造成重大安全隐患。目前,水电枢纽缺陷识别主要依靠人工巡检,存在效率低、风险高等问题。提出一种水电枢纽缺陷识别方法,基于完全自注意力机制构建缺陷识别网络,以提高网络捕捉长距离全局信息的能力和缺陷识别精度。设计2个同尺寸的自注意力编码器分支,通过双分支结构完成不同尺度自注意力计算,从而提取多尺度缺陷特征,增强全局语义表达能力。构建一个基于类别向量的自注意力混合融合模块,并对2条分支的多尺度特征进行融合,以有效应对水电枢纽缺陷图像尺度变化大、形态多样等问题。在四川某水电站枢纽缺陷数据集上的实验结果表明,该方法宏查准率可达98.87%,缺陷识别效果优于SVM、ResNet-50、MobileNet v3等方法。     

Timely autonomous identification of UAV safe landing zones

For many applications such as environmental monitoring in the aftermath of a natural disaster and mountain search-and-rescue, swarms of autonomous Unmanned Aerial Vehicles (UAVs) have the potential to provide a highly versatile and often relatively inexpensive sensing platform. Their ability to operate as an ‘eye-in-the-sky’, processing and relaying real-time colour imagery and other sensor readings facilitate the removal of humans from situations which may be considered dull, dangerous or dirty. However, as with manned aircraft they are likely to encounter errors, the most serious of which may require the UAV to land as quickly and safely as possible. Within this paper we therefore present novel work on autonomously identifying Safe Landing Zones (SLZs) which can be utilised upon occurrence of a safety critical event. Safe Landing Zones are detected and subsequently assigned a safety score either solely using multichannel aerial imagery or, whenever practicable by fusing knowledge in the form of Ordnance Survey (OS) map data with such imagery. Given the real-time nature of the problem we subsequently model two SLZ detection options one of which utilises knowledge enabling the UAV to choose an optimal, viable solution. Results are presented based on colour aerial imagery captured during manned flight demonstrating practical potential in the methods discussed.     

DroneRTEF:development of a novel adaptive framework for railroad track extraction in drone images

Pattern Analysis and Applications - Railroad track health monitoring is a challenging yet important task as it affects the safety of railroad systems. Railroad track extraction presents an...