轨道扣件的图像采集与振动补偿方法

基于车载图像采集系统进行轨道扣件检测是保障列车运行安全的一项关键技术,受运行过程中低频机械振动和外界光照的影响,常规的图像采集检测方法所得图像精度较低。为了解决该问题,提出一种基于机器视觉光源模型的振动补偿方法。通过建立线阵相机、激光光源和扣件区域的照明模型消除外界光照的干扰,并采用频域分析法和多传感数据融合技术,设计补偿算法消除振动信息的干扰。最后,通过搭建采集平台进行实验验证,结果表明该方法可以有效地减弱外界光照对成像质量的影响,并对±5°以内的图像偏移实现了修正。与未采用机器视觉光源模型的方法相比,扣件区域对比度提高10倍以上。     

基于DSP的汽车防眩目监测系统研究

为了解决车辆在夜间行驶时对面车灯给驾驶员造成的眩目问题,提出一种基于DSP的汽车防眩目监测系统,采用TI公司的数据图像处理芯片TM320DM642处理CCD摄像头采集到的强光源图像信息,显示屏采用透明基板的黑白液晶显示屏,利用图像识别的方法,自动跟踪强光源的位置,分析强光源状态,提取强光源的特征参数,用不透光的图像区域自动遮挡强光源的光线,从而实现防眩目的目的.系统经实际运行,性能稳定,防眩目效果明显.     

基于立体视觉的目标检测与轨迹预测研究综述

目前基于立体视觉信息的运动目标识别定位、跟踪及轨迹预测是机器视觉领域的研究热点.通过归纳整理相关文献,从双目立体视觉技术、运动目标检测技术、运动目标轨迹预测技术三个方面对基于立体视觉的运动目标检测及轨迹预测进行了概述,分别阐述了相机标定的常见方法、图像特征提取及立体匹配不同算法的适用场景、各运动目标检测方法的优缺点、常...     

基于长短期记忆网络的高速公路车辆变道轨迹预测模型

高速公路车辆车速、车距、行驶方向等因素都是动态变化的,受外界环境干扰,采集到的目标车辆状态特征数据可能存在噪声,导致车辆变道轨迹预测存在误差,为此提出基于长短期记忆网络的高速公路车辆变道轨迹预测模型,有效预测高速公路车辆变道轨迹,改善车辆行驶条件,保障其安全运行。通过激光雷达、GPS等装置采集目标车辆交通数据,将其合理组合成目标车辆状态观测特征向量,并构建相应的特征向量矩阵,将所构建目标车辆状态观测特征向量矩阵作为1层卷积神经网路输入,提取目标车辆状态观测特征向量潜在特征后,以1层卷积神经网络输出结果为双向长短期记忆网络有效输入,经过无数次模型训练后,输出目标车辆变道轨迹预测结果。实验结果表明:该模型可有效预测高速公路车辆变道轨迹,预测出的轨迹横纵坐标误差极低,能够得到较为理想的高速公路车辆变道轨迹预测结果。     

一种夜晚道路环境下的后方车辆检测方法

目前,基于视觉传感器的车辆检测已经成为车辆驾驶辅助领域的研究热点。但是迄今为止,大多数研究集中在白天好的光照条件,对夜晚条件下的车辆检测研究较少。本文提出了一种基于车灯的夜晚道路环境下的车辆检测算法,利用摄像机采集实时图像来检测自车后方的车辆。首先,基于亮度信息提取夜晚环境图像中的光亮目标物,然后,对提取的光亮目标物进行验证,去除路灯等干扰光源,从而得到真正的车辆头灯;最后,按照基于知识的方法,对提取到的车灯进行组合,并对组合后的车灯对进行验证,从而检测出夜晚道路环境下的车辆。实验结果表明本文算法易于实现,识别率高,适应性好。     

基于机器视觉的驾驶员换道行驶检测研究

为自动检测驾驶员行车路线,通过车载CCD图像传感器获得序列图像,利用计算机视觉技术,建立摄像机的透视投影模型和汽车换道行驶检测模型,应用改进的Hough变换识别道路标线,确立直线方程,采用逆透视投影变换对车辆在当前位置的横向距离和横向偏转角做出估计,进而对汽车行驶状态做出判断。实验证明,该方法能够正确检测和判断驾驶员是否处于压线行驶和换道行驶状态。     

基于机器视觉的火车驾驶员动态手势识别方法

为快速准确识别火车驾驶员动态手势,提出一种基于机器视觉的动态时间规整算法.采用Kinect视觉传感器提取手势深度信息,结合人体骨骼节点信息,通过选取合适的深度距离阈值将手势图像信息从背景中分离出来.基于支持向量机(SVM)算法对分割后的手势图像进行识别并对手势规范性进行评价.利用图像深度数据以及驾驶员骨骼数据得到手臂骨...     

基于机器视觉的喷涂机器人轨迹规划与涂装质量检测研究综述

随着智能喷涂技术的快速发展,机器视觉在喷涂机器人系统中的研究和应用引起广泛关注,合理的喷涂轨迹能保障油漆厚度均匀、减少漆膜缺陷产生,并且融合涂装质量检测技术形成闭环的喷涂系统.鉴于此,针对机器视觉在喷涂机器人轨迹规划与涂装质量检测中的研究进行综述.首先,对喷涂系统在现代产品制造中的快速发展所面临的机遇、挑战和机器视觉技术进行介绍;然后,综述基于机器视觉技术的喷涂机器人轨迹规划和涂装质量检测的研究成果,对基于机器视觉的喷涂机器人轨迹规划方法,包括待喷涂工件的三维重建、基于点云数据的喷涂轨迹自动规划和基于视觉伺服的喷涂轨迹补偿进行分析和讨论,并重点介绍机器视觉在涂装质量检测中的应用与研究现状,从数据增强和模型选择两个方面,对不同任务中基于深度学习的涂装质量检测算法性能的改善提供潜在解决方案;最后,总结与展望机器视觉技术在喷涂机器人轨迹规划与涂装质量检测中的研究方法和思路,为喷涂系统朝着智能化、柔性化的方向发展提供参考.     

视觉SLAM在室内动态场景中的应用研究

视觉SLAM（Simultaneous Localization And Mapping,同时定位与建图）是移动机器人领域的核心技术,传统视觉SLAM还难以适用于高动态场景并且地图中缺少语义信息。提出一种动态环境语义SLAM方法,用深度学习网络对图像进行目标检测,检测动态目标所在区域,对图像进行特征提取并剔除动态物体所在区域的特征点,利用静态的特征点进行位姿计算,对关键帧进行语义分割,在构建语义地图时滤除动态物体的地图点构建出无动态物体干扰的语义地图。在TUM数据集上进行实验,结果显示该方法在动态环境下可以提升88.3%位姿估计精度,并且可同时构建出无动态物体干扰的语义地图。     

基于机器视觉的轴承防尘盖表面缺陷检测

为实现工业现场中轴承防尘盖表面缺陷的自动检测,提出一种基于机器视觉技术的检测方法。采用蓝色同轴光源作为检测系统所用光源,克服金属反光;采用最小二乘法拟合轴承外圆,根据轴承型号比例分割出防尘盖区域,利用Otsu阈值分割和Roberts边缘提取处理图像,每2°统计值为1的点的数目,与模板轴承此数据比较,求出相差角度,由此将防尘盖字符、非字符区域分离,两部分是否存在缺陷分开判别,互不干扰。实际测试表明：检测系统采集到的轴承图像清晰,缺陷检测算法正确率在96%以上,可实现轴承防尘盖表面缺陷的自动检测。     

The effectiveness of glare-reducing glasses on simulated nighttime driving performance in younger and older adults

Glare from oncoming headlights is a problem for nighttime drivers because it can decrease visual acuity and cause discomfort. This diminished visual ability and discomfort due can increase drivers’ risk for traffic accidents. Older drivers experience more severe detrimental effects from nighttime glare, and these effects may pose a growing roadway hazard as the number of older drivers increases. The increased brightness of popular high-intensity-discharge (HID) headlights may further exacerbate these visibility problems. In a sample of younger (under 40 years of age) and older (40 years of age and older) drivers, we examined the impact of headlight glare from HID and traditional halogen lights on driving performance in a simulator, as well as the effectiveness of novel polarized glare-reducing eyeglasses for mitigating glare-induced performance deficits. The glare-reducing glasses increased visual awareness in the face of oncoming HID headlights compared to halogen headlights in both age groups. Older drivers performed significantly worse than did younger drivers on several measures of driving and visual detection performance. The glare-reducing glasses mitigated performance deficits, with older drivers performing similarly to younger drivers when exposed to HID headlights while wearing the polarized glasses. Due to the introduction of brighter LED-based headlights to the consumer automotive market and an expanding population of older drivers, automotive manufactures should consider glare-mitigation strategies when designing future headlight systems.     

地铁车厢灯光自适应控制研究

解决地铁节电问题一直是国内外大力研究的课题。而行驶在高架以及地面段的地铁列车具有天然的采光资源——自然光。根据这一现状,提出一种自适应调节车厢内灯具照度的方法,来达到节约地铁车厢内照明用电的目的。针对自然光随机性高,变化频率快,以及地铁运行时外界环境干扰大的问题,提出采用数字滤波器结合模糊控制的方法,将经巴特沃斯低通滤波器滤除高频干扰的车厢内照度信号传输给模糊控制器,通过模糊控制器的逻辑推理判断处理,实现了地铁车厢灯光自适应地调节车厢内灯具的照度,节约了地铁车厢照明用电。     

基于车前灯的夜间车辆视频检测研究

针对夜间车辆视频检测和车流量统计的难题，提出了一种改进的基于视频图像处理提取车前灯的算法。通过分析夜间车辆视频的特点，利用梯度滤波法消除地面反射光对车灯的干扰，实现图像增强，并将分水岭分割算法和直方图双峰法相结合提取车前灯的信息。利用车灯配对匹配原则设计了一种新的匹配算法和跟踪算法实现车灯的配对与跟踪，最终准确地实现了车辆检测和车流量的统计功能。     

An empirical investigation of a dynamic brake light concept for reduction of rear-end collisions through manipulation of optical looming

The concept of dynamically manipulating the optical looming cues of a lead vehicle's brake lights is investigated as a means of potentially reducing the frequency of rear-end collisions in automobile driving. In a low-fidelity driving simulator, 40 participants were instructed to follow a leading vehicle (LV) and appropriately respond to braking of the LV, under three visibility conditions: day, night-time with following vehicle (FV) headlights, and night-time without FV headlights. During some LV braking events, separation and size of the brake lights of the LV were expanded or contracted, by a nominally imperceptible amount, to simulate an effective virtual time shift in the headway of the LV. Results show that this manipulation was most effective for very poor visibility conditions: at night with no headlights, for which LV brake lights were most salient. When confronting a LV with expanding or contracting brake lights, subjects generally braked sooner or later respectively, in comparison with the no manipulation case. The concept shows some promise for causing drivers to brake sooner in emergencies.     

Transforming car glass into microphones using piezoelectric transducers

The use of car windows as a long range acoustic sensing device for external alarm signals is described in this paper. The goal is to detect and localize siren signals (e.g. ambulances and police cars) and to alert presbycusic drivers of its presence by visual and acoustic feedback in order to improve individual mobility and increase the sense of security. The glass panes of a Renault Zoé operating as an acoustic antenna have been equipped with large 50 mm outer diameter piezoceramic rings, hidden in the lower part of the door structure and the lower part of the windshield and the rear window. The response of the glass to a sweep excitation has been recorded. In general, the glass pane is acting as a high pass filter due to its inherent stiffness and provides only little damping. This effect is compensated by using a charge amplifier electronic circuit. The detection capability up to 120 m in static conditions as well as the influence of wind and vibration during driving conditions is shown. Finally, two alarm detection scenarios are reported with the car passing by a static alarm source once facing the car and once orthogonally to the driving direction.     

基于机器视觉技术的易拉罐罐底喷码检测系统设计

本文根据食品饮料行业易拉罐生产线的工作环境以及罐底喷码检测的检测要求,研制了基于康耐视机器视觉的易拉罐罐底喷码检测系统,实现对易拉罐喷码不合格的产品进行自动检测与快速剔除。该检测系统由光源与视觉处理系统、电气控制与人机交互系统、次品剔除装置等组成。当易拉罐通过成像系统时,金属接近开关触发光源频闪和工业智能相机,获得高速易拉罐罐底图像,智能相机对其分析处理,由电气控制系统执行检测结果,从而达到分拣不良品的目的。通过实际项目应用证明：该系统实时性好,可靠性高,有效地提高了在生产过程中产品喷码质量的控制。     

A vision-centered multi-sensor fusing approach to self-localization and obstacle perception for robotic cars

Most state-of-the-art robotic cars’ perception systems are quite different from the way a human driver understands traffic environments. First, humans assimilate information from the traffic scene mainly through visual perception, while the machine perception of traffic environments needs to fuse information from several different kinds of sensors to meet safety-critical requirements. Second, a robotic car requires nearly 100% correct perception results for its autonomous driving, while an experienced human driver works well with dynamic traffic environments, in which machine perception could easily produce noisy perception results. In this paper, we propose a vision-centered multi-sensor fusing framework for a traffic environment perception approach to autonomous driving, which fuses camera, LIDAR, and GIS information consistently via both geometrical and semantic constraints for efficient self-localization and obstacle perception. We also discuss robust machine vision algorithms that have been successfully integrated with the framework and address multiple levels of machine vision techniques, from collecting training data, efficiently processing sensor data, and extracting low-level features, to higher-level object and environment mapping. The proposed framework has been tested extensively in actual urban scenes with our self-developed robotic cars for eight years. The empirical results validate its robustness and efficiency.     

A Review of:“The Remembering Self: Construction and Accuracy in the Self-narrative”, edited by ULRIC NEISSER and ROBYN FIVUSH,Cambridge University Press,New York (1994), pp. x + 301, £30.00, ISBN 0-521-43194-8

Driving is a task that requires the timely detection of critical events and relevant changes in traffic circumstances. Adaptation of speed and safety margins allows drivers to control the time available to react to potential hazards. One of the basic safety margins in driving is the time headway preserved with respect to cars ahead. To avoid rear-end collisions, drivers have to detect decelerations of lead cars. It can be assumed that fast or abrupt decelerations of the lead car are detected faster than slow or gradual decelerations. Moreover, expected decelerations are presumably detected faster than unexpected decelerations. Drivers' responses to rather abrupt and more gradual decelerations of the lead were investigated in a driving simulator. Situational traffic cues were used to manipulate the driver's expectations. Drivers adjusted the timing of their responses very well to the level of deceleration of the lead car. If cues in the environment indicated that the lead car was likely to decelerate, drivers reacted faster. Moreover, drivers increased their headway before the lead car actually started to decelerate, which can be considered an anticipatory response. In general, anticipation allows drivers to maintain their preferred headway and control time pressure in driving.     

基于麦克纳姆轮的智能车寻迹控制策略研究

为适应复杂工作环境运行需要,解决普通轮式小车在狭小空间内自如转向以及巡线不稳定和路径类型识别不准确的问题,设计了一种以摄像头为视觉传感器,以麦克纳姆轮为驱动的全向智能车.设计以CH32V103R8T6单片机为核心,以MT9V034灰度摄像头为视觉传感器,采用迭代法阈值对图像进行抗干扰灰度处理,通过智能车行驶中视觉道路宽...     

基于OpenCV的机器视觉尺寸测量研究

通过对激光测距传感器与机器视觉的研究利用激光测距传感器、视觉传感器相结合的方法,提出了一种不同于结构光和双目尺寸测量的更为简单有效的新方法和新技术。首先对灰度直方图比较分析选取了最佳滤波方法,然后通过激光测距传感器测距和单目视觉传感器进行边缘的精确检测。激光测距传感器实现了对采样目标车厢进行快速测距定位,计算出采样机械手与目标车辆的相对位置关系。视觉传感器准确识别车厢尺寸及定位,为煤炭采样机械手自动采样提供了实现基础。检测目标是拉煤货运火车车厢,给煤炭采样行业带来更为简单有效,准确的目标定位的识别系统和方法。