基于虚拟线圈的内河船舶流量检测方法

实时检测内河船舶流量对水上交通管理具有重要意义.为实时检测船舶流量,研究了一种基于虚拟线圈的船舶流量检测系统.虚拟线圈即在视频图像上设置一个封闭区域,根据该区域内图像的变化检测是否有运动目标通过.利用RGB三通道背景差分法得到视频图像的二值化图像,二值化图像的三个分割阈值由大津法求出.设置2个平行的虚拟线圈,通过虚拟线圈的船舶会被检测并计数,同时检测船舶的船长与船宽,利用BP神经网络对船舶进行分类.通过在武汉长江大桥和武汉长江二桥上不同时间段采集的视频进行实验,结果表明,船舶计数正确率达到97.1%,计数漏检率2.9%,计数错检率0%,船舶分类正确率98.6%.处理一帧图片的平均时间为7 ms,具有较好的实时性.      

基于虚拟线圈的车型识别研究

在车型识别系统中,车辆的特征提取是非常重要的.传统的检测方法受摄像机的成像精度与架设方式等因素的影响,给车辆特征的提取带来困难.提出了一种采用虚拟线圈识别车型的方法.该方法将检测线与虚拟线圈相结合,对提取的车辆信息进行预处理,以检测线检测到达预先设置位置的车辆,然后触发虚拟线圈来提取车辆特征,通过BP神经网络来训练识别的车辆特征,以达到车型识别的目的.试验证明,该方法识别效果好,识别有效率达99%.     

基于深度置信网络的多源信息前方车辆检测EI北大核心CSCD

本文中以深度置信网络为理论基础,提出了一种多源信息的前方车辆检测方法。首先将毫米波雷达和摄像机进行联合标定,确定两个传感器坐标系之间的转化关系。然后通过对毫米波雷达数据进行预处理完成前方障碍物的标签分类,获得前方车辆目标和其他类障碍物的数据。接着利用深度置信网络对数据进行训练,完成前方车辆的初识别。最终根据常见车型宽度和高度的统计数据获得前方车辆识别的验证窗口。实验结果表明,采用所提出方法前方车辆识别的正确率为91.2%,单帧图像的总处理时间为37ms,有效地提高了系统实时处理速度,尤其对阴天、夜间、轻雨或雾霾等恶劣的道路环境中的车辆有良好的检测效果,能满足汽车辅助驾驶对于准确性和稳定性的要求。     

基于深度置信网络的多源信息前方车辆检测

本文中以深度置信网络为理论基础,提出了一种多源信息的前方车辆检测方法。首先将毫米波雷达和摄像机进行联合标定,确定两个传感器坐标系之间的转化关系。然后通过对毫米波雷达数据进行预处理完成前方障碍物的标签分类,获得前方车辆目标和其他类障碍物的数据。接着利用深度置信网络对数据进行训练,完成前方车辆的初识别。最终根据常见车型宽度和高度的统计数据获得前方车辆识别的验证窗口。实验结果表明,采用所提出方法前方车辆识别的正确率为91.2%,单帧图像的总处理时间为37ms,有效地提高了系统实时处理速度,尤其对阴天、夜间、轻雨或雾霾等恶劣的道路环境中的车辆有良好的检测效果,能满足汽车辅助驾驶对于准确性和稳定性的要求。     

基于倾向流和深度学习的机场运动目标检测

针对当前基于视频图像的场面监视目标检测方法存在定位误差较大,识别准确率低等问题,建立一种结合目标运动信息的机场场面运动目标检测方法:利用倾向流法提取出运动目标在图像中的候选区域,对候选区域执行点池化操作以确定区域建议的边界,采用Inception结构构建一个浅层卷积神经网络,并使用该网络对区域建议中的航空器、车辆和人员进行识别.结合国内机场的监视视频,构建了一个包含4 938张图片的机场目标数据集,用于算法的训练和测试.结果 表明,运动目标提取的准确率达到94%以上,运动目标识别的Top-1准确率达到了97.23%,运动目标平均准确率达到86.23%.与3种深度学习目标检测算法相比,运动目标检测精度平均提升了39%.      

基于视频图像的车辆计数新方法

针对道路车流量检测问题,从便捷性、实时性角度出发,结合视频图像处理技术,对视频车辆计数进行了研究。直接在RGB图像中进行自适应背景更新,以此为基础,对RGB图像进行背景差分,提取出运动车辆区域,避免了复杂环境下图像灰度化过程中的信息丢失;利用当前帧和背景帧的HSI颜色空间信息来滤除阴影;通过在视频图像中设置固定虚拟检测区,实现对车辆的计数。实验结果表明,该方法计算量较小,白天情况下的计数准确率在89.58%以上;夜间的计数准确率较低,还需进一步研究改进。     

基于深度学习的车辆检测及车型识别研究

正当前,在新一代信息技术引领下,数据快速积累、运算能力持续提升、算法模型不断优化、多场景应用快速兴起,人工智能发展环境发生了深刻变化。车辆检测及车型识别作为深度学习目标检测领域在智能交通的重要应用,也是近年来国内外学者的研究热点。本文针对已有的车辆检测方法缺乏车型识别问题,利用深度学习图像识别技术,提出了基于Faster R-CNN的车辆检测及车型识别方法。通过将Faster R-CNN深度学习模型应用     

基于深度学习的车辆目标检测算法综述

车辆目标检测是自动驾驶环境感知的重要组成部分。近年来随着深度学习在目标识别领域取得重大突破,基于深度学习的车辆目标检测算法逐渐成为该领域的研究热点。论文对当前主流的两阶段车辆目标检测算法和单阶段车辆目标检测算法进行简要介绍,分析了其中几种具有代表性的卷积神经网络算法的优缺点,最后总结目前车辆目标检测存在的问题以及未来的发展方向。     

基于机器视觉对交通场景中车辆速度识别研究

通过对YOLOv5机器视觉框架进行二次开发，同时融合DeepSORT追踪算法，实现对桥梁交通车辆时空信息的提取和车辆轨迹的追踪。改进了传统的虚拟线圈法，实现了对车辆速度的测量，避免对传统方法中因检测线圈的像素变化进行阈值的设定，提高了算法的普适性。最后，将算法应用到实际的场景中与测速仪结果进行对比，其中平均误差在1%以内，误差最大值控制在为15%以内。     

基于联合特征和压缩字典学习的车辆类型识别算法

本文中针对车型识别中计算时间长、识别精度低的问题,提出基于联合特征和压缩字典学习的车型识别算法。首先,利用SIFT算法提取车辆原始图像的纹理特征和车辆图像边缘的形状特征,并将其串联,生成更具差别性的联合特征;然后,构建特征字典并进行字典学习,在此过程中,将特征字典分成大小相同的数据块,利用非常稀疏随机投影矩阵,降低样本数据块的维度,通过稀疏编码和字典更新两个阶段,生成最终压缩特征字典;最后,建立稀疏表示分类模型,通过计算待测目标在字典中的最小重构误差,实现车型识别。实验结果表明,该算法能有效提高车型识别准确率和实时性。     

Hierarchical Scheme of Vehicle Detection and Tracking in Nighttime Urban Environment

In this paper, we propose a novel hierarchical scheme for detection and tracking of vehicles using a vehicle-mounted camera in nighttime under urban environment, where a vehicle can be represented by a pair of taillights and various types of lights are commonplace. The proposed scheme, therefore, mainly focuses on devising robust detection and pairing of taillights in spite of their inherent diversity and continuous transformation in appearance. Thus the appearance symmetry, which many conventional methods rely on, for paring is not guaranteed to be available all the times. Each of the three layers in the scheme is devised to identify a vehicle from individual lights and clutters detected in a hierarchical manner. Robust detection of a pair of taillights, which can be regarded as a vehicle, is sought by successive groupings of the components in a layer and checking not only the intra-layer but the inter-layer relations between them. A structural Kalman filter is employed to maintain the temporal consistency in the motion of the components and their relations as well. Exploiting such relational information increases accuracy in tracking of individual components by reducing effects from fluctuation in positions and shapes, and eventually compensating possible failures in detection of them. As a result, the proposed scheme achieves enhancement in detection and tracking of vehicles in nighttime as proven by experiments on videos including crowded urban traffic scenes.     

智能车辆安全辅助驾驶技术研究近况

论述了安全辅助驾驶技术的研究现状、研究的必要性以及研究进展。安全辅助驾驶技术包括车道偏离预警与保持、前方车辆探测及安全车距保持、行人检测、驾驶员行为监测、车辆运动控制与通讯等。分析了各种传感器的优缺点及其在实际应用过程中存在的问题,基于单一传感器不能很好地解决安全辅助驾驶技术可靠性和环境适应能力的要求,应结合激光雷达技术解决图像模糊问题,利用红外传感器增强机器视觉识别的可靠性,未来的安全辅助驾驶技术应该采取多种传感器融合的技术,结合毫米波雷达和激光雷达系统具有深度测量精确的特点,将极大的推动汽车安全辅助驾驶系统的应用和推广。     

基于子波能量和神经网络分类器的机动车车型识别

为了使交通管理系统能进行可靠的机动车分类,研究了轿车、轻型越野车和货车3种机动车目标的声信号,提出了一种采用子波分解后不同尺度上声信号能量作为特征向量的特征提取算法,并设计了kNN(k近邻)分类器和改进BP神经网络分类器用于目标分类。目标识别和分类试验结果表明:所提出的特征提取算法能够很好地体现不同类型目标之间的差异,提取的特征向量稳健;设计的改进BP神经网络分类器的分类精度可达92.6%,且分类效果优于kNN分类器。     

High Performance ACC System Based on Sensor Fusion with Distance Sensor, Image Processing Unit, and Navigation System

Curve sensors used in first generation “Adaptive Cruise Control” systems (ACC) are based on steering angle sensors, lateral accelerometers or yaw rate sensors. The disadvantage of these curve sensors is that they do not have any preview characteristics. This leads in many driving situations to misinterpretations by the ACC system, e.g. wrong path assignments of vehicles ahead because of non-constant curve radii particularly in the beginning and ending of curves. The consequence is that the ACC car brakes due to vehicles in adjacent lanes or it ignores relevant obstacles. In the following a second generation ACC system will be presented whose curve sensor is realized by a real time image processing system with the support of a GPS-based navigation system. This multi-sensor fusion system is now suitable for collision avoidance and stop and go applications.     

Measurement of roof deformation caused by vehicle rollover

A vehicle rollover is a critical accident that could have many causes. This paper describes a novel vision-based system for measuring vehicle roof deformation due to a rollover accident. A vision-based measurement system offers an overall view of structural deformation simply at low cost. Our measurement system was constructed using a Kinect camera from Microsoft, a battery, and a remote-controlled recording computer. Color images and distance maps can be obtained using two sensors embedded in the Kinect along with customized software, and the distance from the camera lens to a specific object can be calculated with a simple equation. To test our proposed approach, actual vehicle rollover experiments were conducted and the resulting roof deformations were compared to those indicated by our system. Moreover, cross-sectional image of Apillar was analyzed to calculate bending moment of inertia. From the research results, it was able to show that deformation errors were within 13 mm, and roof deformation was correlated with vehicle type, or vehicle curb weight.     

Employment of instrumented vehicles to identify real-time snowy weather conditions on freeways using supervised machine learning techniques – A naturalistic driving study

This study introduces the idea of using vehicles as weather sensors to identify real-time weather on freeways in the same context as Road Weather Information System (RWIS) but in a continuous, trajectory-level, and for road segments allocated in the vehicles paths. The study developed a novel approach to detect snowy and clear weather conditions by utilizing real-time data collected from vehicles' external sensors and CANbus. The proposed approach used time series datasets from the SHRP2 Naturalistic Driving Study (NDS), collected during normal driving conditions on freeways. Trips occurring in snowy weather alongside matched trips in clear weather were segmented into time- and distance-based segments such as a one-minute, one-mile, and half a mile. Three assemblies of the input data are considered in the modeling step: data collected from external sensors, CANbus data, and these two data combined. Data analysis was implemented using the Deep Learning Artificial Neural Network, Decision Tree, Random Forest, and Gradient Boosted Trees models. The results indicate that using different segmentation levels provides decent results in detecting snowy weather. The accuracy in estimating the real-time snowy weather was in ranges of 80% to 85%, 71% to 79%, and 73% to 83% for the one-minute, one-mile, and half mile segmentation types, respectively. The GBT model performed the best among all models based on the area under the Receiver Operating Characteristics (ROC) curve, the highest cumulative percentage in estimating the snowy weather using the lower portion of the population, and the highest overall accuracy. Results indicated that an accuracy of 83% in estimating snowy weather conditions could be accomplished using the data collected from external sensors only without accessing CANbus data.     

Two-Dimensional Motion of Four-Wheel Vehicles

This paper deals with two-dimensional motion of vehicles. Since the general four-wheel vehicle model is statically indeterminate, the motion of vehicles has been analyzed by replacing the vehicle with the equivalent two-wheel one proposed by E. Marquard. Because this approximation is based on the axisymmetry of vehicles, it causes significant errors in the general case. To improve the accuracy of the analysis of vehicle motion, a four-wheel model is suggested. In this model, it is assumed that the chassis remains in a flat plane during motion. By introducing this condition, the motion of the vehicle can be analyzed. Several results from the four-wheel model and the equivalent two-wheel model are shown for comparison, and the vehicle trajectories with time are discussed.     

Two-Dimensional Motion of Four-Wheel Vehicles

This paper deals with two-dimensional motion of vehicles. Since the general four-wheel vehicle model is statically indeterminate, the motion of vehicles has been analyzed by replacing the vehicle with the equivalent two-wheel one proposed by E. Marquard. Because this approximation is based on the axisymmetry of vehicles, it causes significant errors in the general case. To improve the accuracy of the analysis of vehicle motion, a four-wheel model is suggested. In this model, it is assumed that the chassis remains in a flat plane during motion. By introducing this condition, the motion of the vehicle can be analyzed. Several results from the four-wheel model and the equivalent two-wheel model are shown for comparison, and the vehicle trajectories with time are discussed.     

用于智能驾驶系统评价的乘员损伤模型

只有较少的交通事故数据资源被用于建立基于碰撞速度信息的乘员损伤模型，致使所得到的模型精度差。为此，提出了基于车辆变形深度的乘员损伤模型。对美国不同制造年代和车辆级别的事故数据进行聚类分析，论证出车辆变形深度与乘员损伤风险具有相关性。以车辆变形深度为自变量，通过回归分析得到乘员损伤模型。不同种类车辆的乘员损伤模型拟合精度R²约为0.9，证明了该模型的正确性。为进一步验证，以此模型为基础，评价智能驾驶系统的有效性。以自动紧急制动系统为例，对比基于变形深度和速度变化量信息2种方法的有效性计算结果。结果表明：2组结果的平均误差不超过1%，验证了基于变形深度的乘员损伤模型的准确性。该模型仅需要事故数据库中准确的变形深度信息，能够获得更多的事故数据支持，从而可以更好地适应于不同类别智能驾驶系统的评价需求。     

Real-time vision-based blind spot warning system: Experiments with motorcycles in daytime/nighttime conditions

This paper describes a real-time vision-based blind spot warning system that has been specially designed for motorcycles detection in both daytime and nighttime conditions. Motorcycles are fast moving and small vehicles that frequently remain unseen to other drivers, mainly in the blind-spot area. In fact, although in recent years the number of fatal accidents has decreased overall, motorcycle accidents have increased by 20%. The risks are primarily linked to the inner characteristics of this mode of travel: motorcycles are fast moving vehicles, light, unstable and fragile. These features make the motorcycle detection problem a difficult but challenging task to be solved from the computer vision point of view. In this paper we present a daytime and nighttime vision-based motorcycle and car detection system in the blind spot area using a single camera installed on the side mirror. On the one hand, daytime vehicle detection is carried out using optical flow features and Support Vector Machine-based (SVM) classification. On the other hand, nighttime vehicle detection is based on head lights detection. The proposed system warns the driver about the presence of vehicles in the blind area, including information about the position and the type of vehicle. Extensive experiments have been carried out in 172 minutes of sequences recorded in real traffic scenarios in both daytime and nighttime conditions, in the context of the Valencia MotoGP Grand Prix 2009.