Inception single shot multi-box detector with affinity propagation clustering and their application in multi-class vehicle counting

Multi-class vehicle detection and counting in video-based traffic surveillance systems with real-time performance and acceptable precision are challenging. This paper proposes a modified single shot multi-box convolutional neural network named Inception-SSD (ISSD) for vehicle detection and a centroid matching algorithm for vehicle counting. An Inception-like block is introduced to replace the extra feature layers in the original SSD to deal with the multi-scale vehicle detection to enhance smaller vehicles’ detection. Non-Maximum Suppression (NMS) is replaced with Affinity Propagation Clustering (APC) to improve the detection of nearby occluded vehicles. For a 300 × 300 input image, on PASCAL VOC 2007 test data set, the proposed ISSD achieved 79.3 mean Average Precision (mAP) and ran on an NVIDIA RTX2080Ti; the network attains a speed of 52.3 frames per second. ISSD with APC generates 2.7% improvement in mAP over original SSD300 while almost retaining its time efficiency. By centroid matching algorithm, the vehicles are counted class-wise with a weighted F1 of 98.5%, which is quite superior to the other recent existing research works.

     

Model based vehicle detection and tracking for autonomous urban driving

Situational awareness is crucial for autonomous driving in urban environments. This paper describes the moving vehicle detection and tracking module that we developed for our autonomous driving robot Junior. The robot won second place in the Urban Grand Challenge, an autonomous driving race organized by the U.S. Government in 2007. The module provides reliable detection and tracking of moving vehicles from a high-speed moving platform using laser range finders. Our approach models both dynamic and geometric properties of the tracked vehicles and estimates them using a single Bayes filter per vehicle. We present the notion of motion evidence, which allows us to overcome the low signal-to-noise ratio that arises during rapid detection of moving vehicles in noisy urban environments. Furthermore, we show how to build consistent and efficient 2D representations out of 3D range data and how to detect poorly visible black vehicles. Experimental validation includes the most challenging conditions presented at the Urban Grand Challenge as well as other urban settings.     

Object tracking method based on data computing

The data computing process is utilized in various areas such as autonomous driving. Autonomous vehicles are intended to detect and track nearby moving objects avoiding collisions and to navigate in complex situations, such as heavy traffic and dense pedestrian areas. Therefore, object tracking is the core technology in the environment perception systems of autonomous vehicles and requires the monitoring of surrounding objects and the prediction of the moving states of objects in real time. In this paper, a multiple object tracking method based on light detection and ranging (LiDAR) data is proposed by using a Kalman filter and data computing process. We suppose that the movements of the tracking objects are captured consecutively as frames; thus, model-based detection and tracking of dynamic objects are possible. A Kalman filter is applied for predicting posterior state of tracking object based on anterior state of the tracking object. State denotes the positions, shapes, and sizes of objects. By computing the likelihood probability between predicted tracking objects and clusters which registered from tracking objects, the data association process of the tracking objects can be generated. Experimental results showed enhanced object tracking performance in a dynamic environment. The average matching probability of the tracking object was greater than 92.9%.

     

基于稀疏帧检测的交通目标跟踪

为了获取高速公路交通视频中目标车辆的行驶轨迹,提出一种基于视频的多目标车辆跟踪及实时轨迹分布算法,为交通管理系统和交通决策提供目标车辆交通信息.首先,使用YOLOv4算法检测目标车辆位置及置信度.其次,在不同场景条件下,使用提出的基于稀疏帧检测的跟踪方法,结合KCF跟踪算法,将车辆数据进行关联获取完整轨迹.最后,用车辆分布图和交通场景俯视图显示轨迹,便于交通管理与分析.实验结果表明,提出的跟踪方法在车辆跟踪中有较高的跟踪正确率,同时基于稀疏帧检测的跟踪方法处理速度也较快,实时轨迹分布正确反映了真实场景的车道信息以及目标车辆运动信息.     

基于改进YOLOv3的快速车辆检测方法

对图像或视频数据中的车辆进行检测是城市交通监控中非常重要并且具有挑战性的任务。该任务的难度在于对复杂场景中相对较小的车辆进行精准地定位和分类。针对这些问题，提出了一个单阶段的深度神经网络（DF-YOLOv3），实现城市交通监控中不同类型车辆的实时检测。DF-YOLOv3对传统的YOLOv3算法进行改进，首先增强深度残差网络提取车辆特征，然后设计6个不同尺度的卷积特征图，并与残差网络中相应尺度的特征图进行融合，形成最终的特征金字塔执行车辆预测任务。在KITTI数据集上的实验表明，提出的DF-YOLOv3方法在精度和速度上均能获得较高的检测性能。具体地，对于512×512分辨率的输入模型，基于英伟达1080Ti GPU，DF-YOLOv3获得93.61%的mAP（均值平均精度），速度达到45.48 f/s（每秒传输帧数）。特别地，对于精度，DF-YOLOv3比Fast R-CNN、Faster R-CNN、DAVE、YOLO、SSD、YOLOv2、YOLOv3与SINet表现更好。     

融合特征的卫星视频车辆单目标跟踪

目的 卫星视频作为新兴遥感数据,可以提供观测区域高分辨率的空间细节信息与丰富的时序变化信息,为交通监测与特定车辆目标跟踪等应用提供了不同于传统视频视角的信息。相较于传统视频数据,卫星视频中的车辆目标分辨率低、尺度小、包含的信息有限。因此,当目标边界不明、存在部分遮挡或者周边环境表观模糊时,现有的目标跟踪器往往存在严重的目标丢失问题。对此,本文提出一种基于特征融合的卫星视频车辆核相关跟踪方法。方法 对车辆目标使用原始像素和方向梯度直方图（histogram of oriented gradient,HOG）方法提取包含互补判别能力的特征,利用核相关目标跟踪器分别得到具备不变性和判别性的响应图;通过响应图融合的方式结合两种特征的互补信息,得到目标位置;使用响应分布指标（response distribution criterion,RDC）判断当前目标特征的稳定性,决定是否更新跟踪器的表征模型。本文使用的相关滤波方法具有计算量小且运算速度快的特点,具备跟踪多个车辆目标的拓展能力。结果 在8个卫星视频序列上与主流的6种相关滤波跟踪器进行比较,实验数据涵盖光照变化、快速转弯、部分遮挡、阴影干扰、道路颜色变化和相似目标临近等情况,使用准确率曲线和成功率曲线的曲线下面积（area under curve,AUC）对车辆跟踪的精度进行评价。结果表明,本文方法较好地均衡了使用不同特征的基础跟踪器（性能排名第2）的判别能力,准确率曲线AUC提高了2.9%,成功率曲线AUC下降了4.1%,成功跟踪车辆目标,不发生丢失,证明了本文方法的先进性和有效性。结论 本文提出的特征融合的卫星视频车辆核相关跟踪方法,均衡了不同特征提取器的互补信息,较好解决了卫星视频中车辆目标信息不足导致的目标丢失问题,提升了精度。     

基于多特征融合的视频交通数据采集方法

提出了一种基于多特征融合的视频交通数据采集方法, 核心思想是: 在图像中设置虚拟线圈, 假设车辆从虚拟线圈上驶过时引起像素变化, 通过识别这种像素变化来检测车辆并估计车速. 与现有技术相比, 本文的贡献在于: 1) 综合利用虚拟线圈内的前景面积、纹理变化、像素运动等特征来检测车辆, 提出了有效的多特征融合方法, 显著提高了车辆检测精度; 2) 根据单个虚拟线圈内的像素运动向量来估计车速, 避免了双线圈测速法的错误匹配问题. 算法测试结果表明本文算法能够在复杂多样的交通场景和天气条件下, 准确地检测车辆和估计车速. 在算法研究的基础上, 研制了一款嵌入式交通视频检测器, 在路口长期采集交通数据, 为交通信号控制和交通规律分析提供决策依据.     

交通道路行驶车辆车标识别算法

为解决交通道路行驶车辆车标识别中存在的目标小、噪声大、种类多的问题,提出了一种基于深度学习的目标检测算法与基于形态学模板匹配算法相结合的方法,并设计了一种高准确度且能应对新类型车标的识别系统.首先,采用通过K-Means++重新聚类锚框值,并引入残差网络的YOLOv4进行车标的一步定位;其次,通过对标准车标图像进行预处...     

TRAFFIC VIDEO-BASED MOVING VEHICLE DETECTION AND TRACKING IN THE COMPLEX ENVIRONMENT

Moving vehicle detection and tracking is the key technology in the intelligent traffic monitoring system. For the shortcomings and deficiencies of the frame-subtraction method, a novel Marr wavelet, kernel-based background modeling method and a background subtraction method based on binary discrete wavelet transforms (BDWT) are introduced. The background model keeps a sample of intensity values for each pixel in the image and uses this sample to estimate the probability density function of the pixel intensity. The density function is estimated using a new Marr wavelet kernel density estimation technique. The background and current frame are transformed by BDWT, and moving vehicles are detected in the binary discrete wavelet transforms domain. For the shortages of RGB (Red, Green, Blue) or HSV (Hue, Saturation, Value) color space-based vehicle shadow segmentation algorithms, shadow segmentation algorithm based on YCbCr color space and edge detection is proposed. The original data of the shadow according to the characteristics of the YCbCr space is chosen, and then, combined with edge detection, the shape and location of the vehicle region is determined. An automatic particle filtering algorithm is used to track the vehicle after detection and obtaining the center of the object. An actual road test shows that the algorithm can effectively remove the influence of pedestrians and cyclists in the complex environment, and can track the moving vehicle exactly. The algorithm with better robustness has a practical value in the field of intelligent traffic monitoring.     

Traffic sign detection based on improved faster R-CNN for autonomous driving

The timely and accurate identification of traffic signs plays a significant role in realizing the autonomous driving of vehicles. However, the size of traffic signs accounts for a low proportion of the input picture, which increases the difficulty of detection. This paper proposes an improved faster R-CNN traffic sign detection method. ResNet50-D feature extractor, attention-guided context feature pyramid network (ACFPN), and AutoAugment technology are designed for the faster R-CNN model. ResNet50-D is selected as the backbone network to obtain more characteristic information. ACFPN is performed to decrease the loss of contextual information. And data augmentation and transfer learning are adopted to make model training more convenient and time-saving. To prove the availability of the proposed method, we compare it with mainstream approaches (SSD, YOLOv3, RetinaNet, cascade R-CNN, FCOS, and CornerNet-Squeeze) and state-of-the-art methods. Experimental results on the CCTSDB dataset show that the improved faster R-CNN achieves the frames per second of 29.8 and the mean average precision of 99.5%, which is superior to the state-of-the-art methods and more suitable for traffic sign detection. Moreover, the proposed model is extended to the Tsinghua-Tencent 100 K (TT100K) dataset and also achieves a competitive detection result.

     

视频中运动目标的实时检测和跟踪

基于视频的自动目标检测和跟踪是计算机视觉中一个重要的研究领域，特别是基于视频的智能车辆监控系统中的运动车辆的检测和跟踪。提出了一种自适应的背景相减法来分割运动物体，为了准确地定位运动车辆的区域，采用差分图像投影和边缘投影相结合的方法来定位车体，同时利用双向加权联合图匹配方法对运动车辆区域进行跟踪，即将对运动车辆区域跟踪问题转化为搜索具有最大权的联合图的问题。该算法不仅能实时地定位和跟踪直道上运动的车辆，同时也能实时地定位和跟踪弯道上运动的车辆，从实验结果看，提出的背景更新算法简单，并且运动车辆区域的定位具有很好的鲁棒性，从统计的检测率和运行时间来看，该算法具有很好的检测效果，同时也能满足基于视频的智能交通监控系统的需要。     

基于改进YOLOv3的高速公路隧道内停车检测方法

为了更准确地检测高速公路隧道内停车行为,提出一种基于改进YOLOv3车辆检测模型的高速公路隧道内停车检测方法。通过筛选VOC数据集以及实际高速公路隧道内的车辆图片制作专门用于高速公路隧道内车辆检测的数据集,选取YOLOv3目标检测模型作为车辆检测的基础网络结构,并对其进行加深网络结构的改进使其能够准确检测隧道内的车辆。将Deep SORT跟踪算法应用于改进的停车检测模型中,对车辆进行跟踪从而计算行驶速度,并创新性地设置双重速度阈值来判别车辆的停车行为。实验结果表明,经过改进的YOLOv3模型相比于原模型,在VOC-vehicle数据集和Tunnel-vehicle数据集上的mAP都有所提升,最终获得了mAP为98.19%的高速公路隧道车辆检测模型。将基于改进YOLOv3的高速公路隧道内停车检测方法在高速公路隧道视频上进行测试,可以有效地在高速公路隧道中完成停车检测的任务。     

背景估计与运动目标检测跟踪

基于视频的自动目标检测和跟踪是计算机视觉中一个重要的研究领域,特别是基于视频的智能车辆监控系统中的运动车辆的检测和跟踪。提出一种自适应的背景估计方法来实时获得当前背景图像,从而分割出运动物体。为了准确地定位运动车辆的区域,采用差分图像投影和边缘投影相结合的方法来定位车体,同时利用双向加权联合图匹配方法对运动车辆区域进行跟踪,即将对运动车辆区域跟踪问题转化为搜索具有最大权的联合图的问题。该算法不仅能实时地定位和跟踪直道上运动的车辆,同时也能实时地定位和跟踪弯道上运动的车辆,从实验结果看,提出的背景更新算法简单,并且运动车辆区域的定位具有很好的鲁棒性,从统计的检测率和运行时间来看,该算法具有很好的检测效果,同时也能满足基于视频的智能交通监控系统的需要。     

基于卷积神经网络的实时车辆检测

为了解决对于尺度变换较大车辆及遮挡车辆检测性能不足的问题,提出了一种实时车辆检测模型.针对车辆检测算法对于尺度敏感的问题,通过使用深度残差网络作为特征提取层,构建特征金字塔网络用于多尺度检测;利用软化非极大抑制线性衰减置信得分解决车辆遮挡问题,从而降低车辆的漏检率;同时对模型进行通道级裁剪缩减模型参数规模,节省计算资源...     

Real-time occlusion tolerant detection of illegally parked vehicles

Illegally parked vehicle detection systems are considered crucial elements in the development of any video-surveillance based traffic-management system. The major challenges in this task lie in making the end solution real time, illumination invariant and occlusion tolerant. A two-stage application framework is presented which efficiently identifies vehicles parked illegally in restricted parking-zones. A real-time approach has been followed and an improved foreground segmentation method based on Segmentation History Images (SHI) is developed to identify stationary objects. A three step pixel based classification method is applied on the background segmentation output to segment adjacent moving pixels that become stationary for certain periods of time. The process then locks on to all identified stationary pixel patches, parts of which overlap with the regions of interest marked interactively a priori. The second stage of the process is applied subsequently to track all the stationary pixel patches detected during the first stage using an adaptive edge orientation based tracking method. Experimental results show that the tracking technique gives more than a 90% detection success rate, even if objects are partially occluded. The technique has been tested on the UK Home Office i-LIDS Parked Vehicle video sequences along with the University of Sussex Traffic Dataset and results are compared with other available state of the art methods.     

基于视频的高速公路车辆提取与跟踪算法研究

将背景重建技术应用于运动车辆的提取与跟踪,构建了一个运动车辆提取与跟踪算法。在引入多特征匹配的基础上,设计了一种简单实用的目标跟踪多特征匹配判决逻辑。实验结果表明,该算法能够完成对多个运动目标的跟踪,而且对光线变化及目标运动状态的变化等不利因素有较好的适应性。     

An efficient and flexible approach for multiple vehicle tracking in the aerial video sequence

ABSTRACT

Multiple vehicle tracking (MVT) in the aerial video sequence can provide useful information for the applications such as traffic flow analysis. It is challenging due to the high requirement for the tracking efficiency and variable number of the vehicles. Furthermore, it is particularly challenging if the vehicles are occluded by the shadow of the trees, buildings, and large vehicles. In this article, an efficient and flexible MVT approach in the aerial video sequence is put forward. First, as the pre-step to approach the MVT problem, the superpixel segmentation-based multiple vehicle detection (MVD) is achieved by combining the two-frame difference and superpixel segmentation. The two-frame difference is utilized to reduce the search space. By scanning the search space via the centre of the superpixels, the moving vehicles are detected efficiently. Then, the deterministic data association is proposed to tackle the MVT problem. To improve the tracking accuracy, we fuse multiple types of features to establish the cost function. With respect to the variable number of the vehicles, the tracker management is designed by establishing or deleting the trackers. Furthermore, for the occlusion handling, we focus on the accurate state estimation, and it is realized by the driver behaviour-based Kalman filter (DBKF) method. In the DBKF method, we take seriously into account the driver behaviour, including the speed limit and rear-end collision avoidance with the front vehicle. Both tracker management and occlusion handling make the MVT approach flexibly cope with varieties of traffic scenes. Finally, comprehensive experiments on the DARPA VIVID data set and KIT AIS data set demonstrate that the proposed MVT algorithm can generate satisfactory and superior results.     

Online multiple objects tracking with detection reliability prior constraint

Multi-object tracking (MOT) is one popular topic in computer vision. It remains a challenging problem in complex scenes, especially of objects with similar appearance. In this case, many existing data association strategies, which link detections among consecutive frames according appearance and motion cues, may fail to track due to unreliable detections or confused appearance and motion. To solve this problem, this paper proposed a novel online multi-object tracking method with detection reliability prior constraint. Our method integrates the trajectory estimation and detection-prediction association into a unified framework. The detection reliability prior constraint is built with the Hankel matrix from object motion model. When we build the Hankel matrix, we adaptively select a set of previous frames to predict object states and calculate the associated weights between detections and candidate objects. Data association in MOT then is estimated by maximum a posteriori (MAP) in a Bayesian framework, accompanied with both previous trajectory and the current detection reliability. Experimental results using synthetic dataset and four public challenging datasets demonstrate that, the proposed method has a good tracking performance compared with the state-of-the-art multi-object trackers.     

一种新的随机游走多车辆检测算法

针对交通视频序列中的多车辆检测问题,提出了基于边缘检测的随机游走算法对多车辆视频进行精确车辆检测。首先利用背景差分结合边缘信息来检测运动车辆区域,针对粘连车辆问题提出了利用检测车辆区域的骨架结构自动完成有效的标记点的提取,把像素标记点作为随机游走算法中目标的种子点;利用种子点对多车辆进行随机游走分割,实现车辆自动检测。