基于立体视觉与光流融合的运动目标检测

针对摄像机与被检测目标同时运动时的目标检测问题,提出一种立体视觉与光流融合的运动目标检测算法。结合立体视觉技术设计了光流与自运动估计模型,运用车辆的运动信息和场景的深度信息估计因摄像机运动产生的自运动光流;采用多分辨率细化的Horn算法估计场景的混合光流;对混合光流和自运动光流进行差分运算,剔除背景中静态目标的运动干扰。经过一系列形态学滤波处理获得运动目标完整区域,依据光流的连通性对运动目标标号,并确定位置信息。以典型的交通场景为对象进行分析,实验结果表明该算法能有效地检测出动态背景下的运动目标。     

复杂地面背景下的红外目标检测算法

提出一种静态场景下的基于帧差光流的随机采样均值漂移聚类检测算法.该方法首先通过隔帧差分法提取运动目标区域,并对运动区域进行光流计算,采用自适应光流阈值分割法准确提取出运动目标;然后运用连通区域标记算法对运动区域进行初步划分,得到若干个连通域子集特征向量样本点,通过提出的随机采样策略来确定对子集空间中样本点的抽样次数;最后利用均值漂移算法对每个子集中的样本点进行若干次抽样计算并分析聚类收敛结果是否相同,从而完成对连通域标记结果的检验.该策略通过减少对标记结果所有样本点的采样次数,提高了目标的检测速度和精度,在不同红外测试场景中的实验结果表明:与传统红外多目标检测算法相比,该方法具有良好的局部抗遮挡性、准确性和实时性,并且检测率能达到95.27%,每帧处理时间达到39.12ms,满足实时处理需求.     

动态场景下基于运动物体检测的立体视觉里程计

为了提升动态场景中视觉里程计的稳健性和精确度,提出一种基于运动物体检测的立体视觉里程计算法。首先,建立考虑相机位姿的场景流计算模型,用于表示物体的运动矢量。其次,提出构造虚拟地图点的方法,一方面结合场景流进行运动物体检测,另一方面使运动物体在图像中占比较大时仍有足够匹配点对用于位姿估计。最后,通过局部地图点及虚拟地图点与当前帧特征点的匹配结果,构建考虑虚拟点的非线性优化模型进行相机位姿估计,既保证静态地图点不与运动物体的特征点形成错误匹配,又避免因有效匹配点对过少而导致视觉里程计失效。数据集实验和实际场景在线实验结果表明,本文算法提升了视觉里程计在动态场景中的稳健性和精确度。     

融合ViBe与帧差法的交叉路口多车辆检测方法

前景检测的精确程度是交叉路口车辆检测的重要因素,传统的基于背景建模的前景检测方法存在拖影现象,并且通常难以分辨出无牌车辆,针对上述问题,提出一种融合ViBe与帧差法的前景检测算法,并在此基础上结合车牌检测算法来检测场景当中的车辆。首先,结合帧差法和ViBe算法对背景像素点的判定结果,采用不同更新因子更新背景模型,其次,使用一种多条件过滤车牌检测算法定位运动区域当中车牌,最后,以检测到的车牌中心为锚点,定位出最终车牌区域。实验结果表明,该前景检测算法可有效应对交叉路口场景下的前景检测的拖影现象,同时车辆检测算法可以准确检测出进入场景时的车辆,并分辨出无牌车辆。     

基于局部光流约束的角点匹配算法

提出了一种基于局部光流约束的角点匹配算法。首先采用Harris算子获得当前帧和参考帧的角点,然后以角点的光流特征作为约束条件,根据两帧图像角点集的坐标分布,排除异常角点,完成角点的精确匹配,实现图像之间的高精度运动估计。通过对视频序列进行实验,采用差图法可主观地发现该运动估计算法的有效性;以峰值信噪比作为评价指标,发现原始视频序列的帧间峰值信噪比明显低于仿射视频序列的帧间峰值信噪比。前者的平均值为22.8072,后者的平均值为33.3854,从而客观地说明了该算法的有效性和稳定性。     

基于稀疏光流法的ORB特征匹配优化

针对图像特征误匹配数量大的问题，提出一种基于稀疏光流法的ORB图像特征点匹配算法。对特征点进行暴力匹配得到初始匹配点集，利用稀疏光流法计算特征点运动向量，估计出特征点在待匹配图像中的二维坐标位置，剔除偏离估计位置较远的特征点匹配对，最后利用随机抽样一致算法进行几何校验进一步优化匹配结果，达到剔除误匹配的效果。实验结果表明:该算法相较于ORB算子、SIFT算子、SURF算子准确率平均提升了21.6%，较RANSAC-ORB算法准确率平均提升了2%，且该算法对图像光照变换、视角变换、模糊变换、旋转和缩放变换和光照变化具有较好的通用性。     

一种快速的单目移动相机下运动目标检测方法

从射影几何的角度分析了单目移动相机下场景运动矢量与摄像机运动之间的关系,基于摄像机光心坐标系,提出了一种快速极线估计算法.该算法中摄像机在此坐标系下永远静止,只有场景和运动目标在运动,将原来移动平台下运动目标检测的问题转换成静止平台下场景全局运动与运动目标独立运动的问题,并推导出光流约束的简洁形式.该算法框架能够根据KLT算法获得Harris角点光流场,并根据实际图像的运动场补偿摄像机的随机运动,同时在保证算法准确性与鲁棒性的前提下,与原来算法相比,计算速度提升了10倍左右.根据实际采集的图像序列进行了分析对比,真实的数据测试表明快速极线估计算法在保证算法准确性与鲁棒性的前提下,极大地降低了算法的计算量与计算时间,从而无需三维重建便可有效地解决单目移动摄像机下运动目标检测的问题.     

抗前景干扰的自适应电子稳像算法

针对场景中存在前景目标运动的抖动视频,提出一种抗前景干扰的自适应电子稳像算法.算法以视觉对运动的感知为指导,采用基于块的三帧间差分,利用时空一致性快速剔除运动前景区域;改进传统Harris算子,用网格筛选和显著度排序,对背景区域进行全局显著特征点的提取和配准,保证全局配准准确度;用统计分布的距离准则去除误匹配点,无需特征点迭代运算,提高了全局运动估计的速度和准确度.在Sage-Husa自适应运动滤波方法的基础上,改进了修正过程噪音和观测噪音的统计特性,模拟摄像机低频匀速运动的视觉平滑效果,有效解决摄像机抖动中存在的扫描运动.在Intel酷睿2四核2.33GHz的微机上用VC++进行实验,结果表明,该算法对320×240像素的视频序列能够达到22fps的处理能力,可以实时稳定含较大或多运动前景目标的复杂抖动视频,输出视觉完整流畅的真实扫描场景.     

基于特征匹配和校验的鲁棒实时电子稳像

提出了一种基于特征点匹配和校验的鲁棒实时电子稳像算法.首先利用Kanade-Lucas-Tomasi角点检测器提取参考帧和当前帧的特征点,并用绝对误差和准则进行特征点匹配|在校验阶段,提出一种能够有效剔除前景运动物体特征点和错误匹配点的空间位置不变准则|最后,在相似运动模型下,利用最小二乘法求解全局运动矢量进行运动补偿.实验证明:该算法满足实时性要求,对视频的平移、旋转、缩放运动都有较好的稳像效果,并对运动物体具有鲁棒性.     

基于聚类的相位相关运动估计方法

针对相机运动引起的图像序列运动的问题,提出了一种基于聚类的相位相关块匹配运动估计算法。利用Harris算子分别在相邻帧图像上检测角点,以参考图像角点为中心选取一个矩形块,将块匹配法与相位相关相结合来计算图像间的运动矢量。最后,对获得的多个块的平移量,进行空间聚类从而选取运动估计比较准确的点。实验结果表明：该算法配准精度能达到亚像素,稳定性较好。     

Vehicle speed measurement based on gray constraint optical flow algorithm

Vehicle speed measurement (VSM) based on video images represents the development direction of speed measurement in the intelligent transportation systems (ITS). This paper presents a novel vehicle speed measurement method, which contains the improved three-frame difference algorithm and the proposed gray constraint optical flow algorithm. By the improved three-frame difference algorithm, the contour of moving vehicles can be detected exactly. Through the proposed gray constraint optical flow algorithm, the vehicle contour's optical flow value, which is the speed (pixels/s) of the vehicle in the image, can be computed accurately. Then, the velocity (km/h) of the vehicles is calculated by the optical flow value of the vehicle's contour and the corresponding ratio of the image pixels to the width of the road. The method can yield a better optical flow field by reducing the influence of changing lighting and shadow. Besides, it can reduce computation obviously, since it only calculates the moving target contour's optical flow value. Experimental comparisons between the method and other VSM methods show that the proposed approach has a satisfactory estimate of vehicle speed.     

动态背景下基于光流场分析的运动目标检测算法

针对现有动态背景下运动目标检测算法的不足,提出一种基于光流场分析的运动目标检测算法.首先根据前背景在光流梯度幅值和光流矢量方向上的差异确定目标的大致边界,然后通过点在多边形内部原理获得边界内部的稀疏像素点,最后以超像素为节点,利用混合高斯模型拟合的表观信息和超像素的时空邻域关系构建马尔可夫随机场模型的能量函数,并通过使目标函数能量最小化得到最终的运动目标检测结果.该算法不需要任何先验假设,能够同时处理动态背景和静态背景两种情况.多组实验结果表明,本文算法在检测的准确性和处理速度上均优于现有算法.     

基于特征匹配和校验的鲁棒实时电子稳像

提出了一种基于特征点匹配和校验的鲁棒实时电子稳像算法.首先利用Kanade-LucasTomasi角点检测器提取参考帧和当前帧的特征点,并用绝对误差和准则进行特征点匹配;在校验阶段,提出一种能够有效剔除前景运动物体特征点和错误匹配点的空间位置不变准则;最后,在相似运动模型下,利用最小二乘法求解全局运动矢量进行运动补偿....     

Stability of vehicles moving on an elastic foundation

Few studies have been made of the stability of a dynamic system moving on an infinite continuum. Here a general method of analysis of such coupled systems is presented. It shows that vehicles possessing a single point of contact with the foundation become unstable above a velocity always higher than the critical speed defined in the classical constant moving force problem. Flutter speeds lower than this critical speed have been obtained in the case of vehicles with two points of contact. This destabilizing effect is due to the damping of the foundation. The evolution of the flutter boundaries as a function of the characteristics of the foundation is described for a typical vehicle.     

Human detection based on motion object extraction and head–shoulder feature

Aimed at the shortcomings of the traditional video monitoring system, human detection method in intelligent video monitoring system was researched. This paper proposed a human detection method based on motion object extraction and head–shoulder feature to complete human detection and statistics in video image sequences. Firstly, background subtraction based on adaptive threshold was used to extract foreground moving object information, then image erosion and image dilation were used to bypass the object shade and remove false object in order to optimize the results of motion object extraction. And finally, for realizing human moving object detection, we proposed the object discrimination algorithm based on human head–shoulder feature to complete human detection and statistics. Experimental results show that the method can successfully realize human detection and statistics. The method is highly accurate and has good real-time and extensive applications. The identification rate is 86% through human video sequences to test. This method can detect human automatically and provide the theoretical and technological base for object detection in the intelligent surveillance system.     

基于深度学习的无人车夜视图像语义分割

为了增强无人车对夜视图像的场景理解,在夜间模式下更快更精确地探测和识别周围环境,将深度学习应用于夜视图像的场景语义分割,提出了一种基于卷积-反卷积神经网络的无人车夜视图像语义分割方法。在传统的卷积神经网络中加入反卷积网络,构建卷积-反卷积神经网络,无需手工选取特征。通过像素到像素的学习和训练,得到图像语义分割模型,可直接用该模型预测夜视图像中每个像素所属的场景语义类别,实现无人车夜间行驶时的环境感知。实验结果表明,该方法具有较好的准确性和实时性,平均IU达到68.47。     

鲁棒的车载红外视频电子稳像算法

针对车辆行进中,红外热像仪拍摄的视频序列存在复杂的随机抖动,提出基于BRISK特征点匹配的运动估计算法,计算出高精度全局运动矢量,同时对于特征点匹配时出现误匹配及场景中存在前景运动物体的情况,采用模糊聚类法分离全局运动和局部运动,提高了算法的鲁棒性。提出了基于Kalman粒子滤波算法,有效实现了复杂扫描运动和随机抖动的分离,并利用双线性插值法进行图像补偿。采用快速图像拼接法进行未定义区域处理,实现了图像全景输出。还利用车载红外热像仪实际拍摄的红外视频进行了稳像实验。实验结果表明,视频序列获得了很好的稳像效果,能够满足实际应用要求。     

Development and calibration of acoustic video camera system for moving vehicles

In this paper, a new acoustic video camera system is developed and its calibration method is established. This system is built based on binocular vision and acoustical holography technology. With binocular vision method, the spatial distance between the microphone array and the moving vehicles is obtained, and the sound reconstruction plane can be established closely to the moving vehicle surface automatically. Then the sound video is regenerated closely to the moving vehicles accurately by acoustic holography method. With this system, the moving and stationary sound sources are treated differently and automatically, which makes the sound visualization of moving vehicles much quicker, more intuitively, and accurately. To verify this system, experiments for a stationary speaker and a non-stationary speaker are carried out. Further verification experiments for outdoor moving vehicle are also conducted. Successful video visualization results not only confirm the validity of the system but also suggest that this system can be a potential useful tool in vehicle's noise identification because it allows the users to find out the noise sources by the videos easily. We believe the newly developed system will be of great potential in moving vehicles' noise identification and control.     

Moving object detection based on improved VIBE and graph cut optimization

Extracting foreground moving objects from video sequences is an important task and also a hot topic in computer vision and image processing. Segmentation results can be used in many object-based video applications such as object-based video coding, content-based video retrieval, intelligent video surveillance and video-based human–computer interaction. In this paper, we present a novel moving object detection method based on improved VIBE and graph cut method from monocular video sequences. Firstly, perform moving object detection for the current frame based on improved VIBE method to extract the background and foreground information; then obtain the clusters of foreground and background respectively using mean shift clustering on the background and foreground information; Third, initialize the S/T Network with corresponding image pixels as nodes (except S/T node); calculate the data and smoothness term of graph; finally, use max flow/minimum cut to segmentation S/T network to extract the motion objects. Experimental results on indoor and outdoor videos demonstrate the efficiency of our proposed method.