面向车辆盲区防撞系统的交通目标智能检测

交通目标智能检测是车辆盲区智能防撞系统中的基础技术,该技术的研究和应用对降低交通事故损失具有重要意义.本文面向车辆盲区防撞系统设计的交通目标智能检测,其在基础模型中融合了两个性能提升策略.将该模型应用于国内和国外道路场景检测数据集,以验证模型在所有范围和近距离目标的检测性能.实验结果表明该模型可以对近距目标表现出较高的...     

基于嵌入式车流量实时检测算法研究与实现

针对道路现场实时车流量检测问题,提出了一种改进的帧间差分法的运动车辆检测算法,并将该检测算法成功移植到了嵌入式系统上。将帧间差分法与采用长度、宽度、面积筛选轮廓及用质心距离的车辆跟踪算法结合,实现运动车辆的检测;将U-Boot引导程序、Linux内核、Yaffs2文件系统和检测算法移植到S3C6410上,通过摄像头实时采集交通视频,检测结果由触摸屏显示。复杂交通场景的实时测试结果表明,本系统的检测时间为0.298秒/帧,准确率超过88%,基本能够实现在道路现场的车流量实时检测。     

基于D-LinkNet与形态学计算的高分遥感图像车流量监测

针对城市交通路面车流量监测实时性和准确性高的需求,设计了一种利用高分遥感图像进行道路网自动提取和车辆自动监测的处理方法。综合利用D-LinkNet和形态学计算实现道路区域的二值化语义分割及连贯性、边缘扩展以及平滑性优化,同时将道路信息作为掩码并外溢后进行车辆目标检测,有效避免非道路区域车辆目标干扰。提出采用热力图的形式改进车流量监测方法,可以更直观显示道路拥挤程度。综合利用现有数据集对所提方法进行评价,车辆检测的平均精度达91.7%,道路提取平均交并比达85.3%,可以实现道路车流量的有效监测。     

Real-time image processing approach to measure traffic queueparameters

The real-time measurement of various traffic parameters including queue parameters is required in many traffic situations such as accident and congestion monitoring and adjusting the timings of the traffic lights. In case of the queue detection, at least two algorithms have been proposed by previous researchers. Those algorithms are used for queue detection and are unable to measure queue parameters. The authors propose a method based on applying the combination of noise insensitive and simple algorithms on a number of sub-profiles (a one-pixel-wide key-region) along the road. The proposed queue detection algorithm consists of motion detection and vehicle detection operations, both based on extracting edges of the scene, to reduce the effects of variation of lighting conditions. To reduce the computation time, the motion detection operation continuously operates on all the sub-profiles, but the vehicle detection is only applied to the tail of the queue. The proposed algorithms have been implemented on an 80386-based microcomputer system and the whole system works in real-time     

基于可变形卷积神经网络的遥感影像密集区域车辆检测方法

车辆检测是遥感图像分析领域的热点研究内容之一,车辆目标的智能提取和识别,对于交通管理、城市建设有重要意义。在遥感领域中,现有基于卷积神经网络的车辆检测方法存在实现过程复杂并且对于车辆密集区域检测效果不理想的缺陷。针对上述问题,该文提出基于端到端的神经网络模型DF-RCNN以提高车辆密集区域的检测精度。首先,在特征提取阶段,DF-RCNN模型将深浅层特征图的分辨率统一并融合;其次,DF-RCNN模型结合可变形卷积和可变形感兴趣区池化模块,通过加入少量的参数和计算量以学习目标的几何形变。实验结果表明,该文提出的模型针对密集区域的车辆目标具有较好的检测性能。     

基于无线通信的山路盲区弯道会车提示系统设计

文中设计的山路拐弯盲区会车提醒系统设置在公路两边,一部分是装载在车辆上。利用单片机、红外发射器和红外接收器构成车辆探测模块,采用无线通信的nRF905单片射频收发器,传输信息到车载路况提示系统上,ISD 语音芯片进行语音播报,液晶显示模块显示数据信息。通过该系统,驾驶员可获得相关路况信息以进行预判,达到降低发生事故、确保山路行车安全的效果。     

Obstacle detection in a road scene based on motion analysis

This paper deals with the problem of obstacle detection in traffic applications. The proposed device allows a driver to receive the current road and vehicle environment information. The perception of the environment is performed through a fast processing of image sequences acquired from a single camera mounted on a vehicle. This approach is based on frame motion analysis. The road motion is first computed through a fast and robust wavelets analysis. Finally, we detect the areas that have a different motion thanks to a Bayesian modelization. Results shown in this paper prove that the proposed method permits the detection of any obstacle on all type of road in various image conditions.     

Vehicle detection by means of stereo vision-based obstacles features extraction and monocular pattern analysis.

This paper presents a stereo vision system for the detection and distance computation of a preceding vehicle. It is divided in two major steps. Initially, a stereo vision-based algorithm is used to extract relevant three-dimensional (3-D) features in the scene, these features are investigated further in order to select the ones that belong to vertical objects only and not to the road or background. These 3-D vertical features are then used as a starting point for preceding vehicle detection; by using a symmetry operator, a match against a simplified model of a rear vehicle's shape is performed using a monocular vision-based approach that allows the identification of a preceding vehicle. In addition, using the 3-D information previously extracted, an accurate distance computation is performed.     

Vision-based road-traffic monitoring sensor

Current techniques for road-traffic monitoring rely on sensors that have limited capabilities and are often both costly and disruptive to install. The use of video cameras (many of which are already installed to survey road networks), coupled with computer vision techniques, offers an attractive alternative to current sensors. Vision-based sensors have the potential to measure a greater variety of traffic parameters (e.g. entry/exit statistic, journey times and incident detection) while installation and maintenance may be performed without disruption to traffic flow. Work on a model based approach for locating vehicles in images of complex road scenes is presented. The location of the vehicle in the image is transformed to the vehicle's position and orientation in the real world while the deformable vehicle model allows the vehicle's principal dimensions to be measured. This data may be passed to a high level tracking algorithm to extract traffic parameters such as vehicle speed, vehicle count, and junction entry/exit statistics. The principal dimensions may be used to classify the vehicle within categories such as car, van or bus. The system could also be used as a boot-strap process for faster, but perhaps less robust, tracking algorithms. The key features of the system are described and results from testing it on images from real traffic scenes are presented     

Multiple mix zones de-correlation trajectory privacy model for road network

Preserving privacy of vehicle movement is an important challenge in road networks; as trajectory data with spatiotemporal information may reveal much individual information. One of the main threats is revealing history location of vehicle trajectories while it stops and again moves toward the destination. Generally, vehicles stop at mostly two places; the first one is traffic light (signal system)/traffic jam and second is at parking locations such as office, shopping mall, home, hospital etc. While existing works only consider social spots. To cope with this issue, we present a new multiple mix zones de-correlation privacy model in which the degree of de-correlation between parking locations and traffic light/traffic jam places. Further, we consider multiple mix zones method to replace parking locations and traffic light/traffic jam places by de-correlation mix zone region. This paper presents an improved privacy traffic monitoring system for road network applications via a proposed security scheme. Specifically, the proposed model analyzes the monitored scene and deployed mix zones parking location and traffic light/traffic jam places. Our method achieved a high privacy level and anonymity solution for trajectory model; moreover, it also balances the service quality and privacy protection. Finally, we performed experiments on real-world data and showed the effectiveness of our method in comparison to existing methods.

     

基于背景模型的车辆目标检测

为了实现对交通车辆快速准确地统计,文中提出一种自适应权值的背景更新方法以适应道路环境的复杂变化.首先在多个通道建立单高斯背景统计模型,然后利用场景中像素的概率分布实现对运动区域的准确检测,最后根据检测结果实现对交通流量的统计.实验结果表明:该方法能够对运动车辆进行快速准确地检测和统计,并对场景的光照变化等影响具有较高的鲁棒性.     

基于视觉的道路场景建模

步入21世纪,随着人工智能的发展,智能车的研究成为一大热点,而智能汽车研究的基础就是定位问题。目前有2种定位方法,一种是实时定位与建图(SLAM),另一种则是基于道路场景表征建模的定位方法,两者各有所长。本文针对基于视觉的道路场景表征建模定位方法进行了优化与改进。首先,本文提出了一种对点云处理的方法,对当前Z坐标一定距离内的点云取不同权值,进行加权投影,以此来构建道路的二维场景。采用ORB特征提取算子提取二维特征,并采用视觉里程计算法获取车辆运动轨迹信息。构建了轻量级神经网络,用来检测道路标志特征,例如车道线、斑马线、道路标志牌等。对二维场景精度差的问题进行补充。     

基于YOLO算法的行人检测方法

针对复杂道路交通环境,选择YOLO(You Only Look Once)实时目标检测算法,对行人目标进行检测识别的研究。YOLO算法在目标检测的速度和精度上都取得过良好效果。首先在YOLO网络模型的基础上针对行人单类检测问题,修改分类器,并通过卷积操作改变网络最后的输出维度;其次通过对道路交通场景下采集到的样本图片进行标注,得到行人数据集;然后采用相同预训练模型在YOLOv2和YOLOv3上训练,通过优化网络参数,加速模型收敛。实验结果分析可知,基于改进的YOLOv3的行人目标检测方法更能满足实时性的要求。     

多特征融合的车辆识别技术

针对运动车辆目标识别问题提出了一种自然场景下车辆识别方法。首先采用图像差分技术对目标车辆的显著特征进行统计学习,并将学习所得目标局部特征以及图像进行编码,根据以上两个信息实现目标车辆的显著性检测。其次针对车辆运动的复杂性,采用分块投影匹配方法进行全局运动估计和补偿,并利用差分技术进行运动特征检测。然后将目标车辆的显著性特征与运动特征进行融合,从而获得更精确的候选目标区域。最后对候选区域进一步使用视觉显著特征进行目标判别。实验表明该方法具有较好的目标判别性能,能较好地解决自然场景下运动车辆的识别问题。     

Location prediction algorithm for a nonlinear vehicular movement in VANET using extended Kalman filter

Vehicular ad-hoc network (VANET) is an essential component of the intelligent transportation system, that facilitates the road transportation by giving a prior alert on traffic condition, collision detection warning, automatic parking and cruise control using vehicle to vehicle (V2V) and vehicle to roadside unit (V2R) communication. The accuracy of location prediction of the vehicle is a prime concern in VANET which enhances the application performance such as automatic parking, cooperative driving, routing etc. to give some examples. Generally, in a developed country, vehicle speed varies between 0 and 60 km/h in a city due to traffic rules, driving skills and traffic density. Likewise, the movement of the vehicle with steady speed is highly impractical. Subsequently, the relationship between time and speed to reach the destination is nonlinear. With reference to the previous work on location prediction in VANET, nonlinear movement of the vehicle was not considered. Thus, a location prediction algorithm should be designed by considering nonlinear movement. This paper proposes a location prediction algorithm for a nonlinear vehicular movement using extended Kalman filter (EKF). EKF is more appropriate contrasted with the Kalman filter (KF), as it is designed to work with the nonlinear system. The proposed prediction algorithm performance is measured with the real and model based mobility traces for the city and highway scenarios. Also, EKF based prediction performance is compared with KF based prediction on average Euclidean distance error (AEDE), distance error (DE), root mean square error (RMSE) and velocity error (VE).     

一种新型多车道车流量检测算法

为了实时有效地检测道路路口车流量信息,并为交通控制和管理提供准确的交通流数据,提出了一种新型的车流量检测算法。通过利用现有的路面标记进行图像对称分割并计算图像灰度值方差,来判断有无车辆通过,进而实现车流量计算。仿真结果表明,该算法不仅简单,易于实现,而且检测准确率高,实时性好,能够有效地为智能交通灯控制提供信息数据。     

基于车联网大数据分析的实时路况检测系统

为解决城市交通拥堵问题,给人们提供优质的出行体验,文中提出了基于车联网大数据分析的实时路况检测系统。使用GPS技术对行驶的车辆进行数据采集,通过数据清洗和数据修复得到样本集合,利用改进模糊C均值聚类算法对样本数据进行分析,得出各路段的平均车速,进而得到相应路段的交通状态。测试结果表明,该系统能够准确得获取道路上行驶车辆的交通数据,识别出当前路段的交通状态,从而证明了该系统设计的合理性和正确性。     

A real-time detector for parked vehicles based on hybrid background modeling

In this paper, a real-time detection system based on hybrid background modeling is proposed for detecting parked vehicles along the side of a road. The hybrid background model consists of three components: (1) a scene background model, (2) a computed restricted area map, and (3) a dynamic threshold curve for vehicles. By exploiting the motion information of normal activity in the scene, we propose a hybrid background model that determines the location of the road, estimates the roadside and generates the adaptive threshold of the vehicle size. The system triggers a notification when a large vehicle-like foreground object has been stationary for more than a pre-set number of video frames (or time). The proposed method is tested on the AVSS 2007 PV dataset. The results are satisfactory compared to other state-of-the-art methods.     

Traffic accident prediction using 3-D model-based vehicle tracking

Intelligent visual surveillance for road vehicles is the key to developing autonomous intelligent traffic systems. Recently, traffic incident detection employing computer vision and image processing has attracted much attention. In this paper, a probabilistic model for predicting traffic accidents using three-dimensional (3-D) model-based vehicle tracking is proposed. Sample data including motion trajectories are first obtained by 3-D model-based vehicle tracking. A fuzzy self-organizing neural network algorithm is then applied to learn activity patterns from the sample trajectories. Finally, vehicle activity is predicted by locating and matching each partial trajectory with the learned activity patterns, and the occurrence probability of a traffic accident is determined. Experiments show the effectiveness of the proposed algorithms.     

汽车超载监测系统

为了实时识别各种车型的超载车辆,该系统基于开源计算机视觉库(OpenCV),先根据车辆照片库建立车型分类器,然后使用数字摄像机拍摄进入监控区域的车辆,在视频中使用分类器识别车型,根据所识别得到的车型去查询数据库获得该车型的核载,再通过动态称重技术获得车辆的实际载重,及时判别车辆是否超载。此方法可避免过去使用统一重量衡量不同车型是否超载的弊端,并可同时免线圈测量车速。测试结果表明系统能快速准确地识别出车型。配合动态称重系统,就能实时得出所通过的车辆是否超载,对公路养护和道路交通安全有相当大的实用意义。