基于双向二维主成分分析的交通标志识别

针对交通标志识别实时性不足,提出了一种基于双向二维的主成分分析[(2D)2PCA]的交通标志识别算法.首先,对交通标志图像进行去噪归一化等预处理.然后,进行水平和垂直方向的投影,通过特征空间降维提高匹配速度.最后,利用最近邻法进行分类.通过在不同数据库下与传统2DPCA方法的对比仿真表明,2种方法随主特征数目增加,识别率都有所提升;样本数量增加时,(2D)2PCA算法的时间增长速度明显小于2DPCA,满足了识别的实时性要求.     

道路标线图像采集中的几何畸变校正研究

为探索道路标线养护信息采集智能化，建立了基于机器视觉的道路标线图像信息采集系统，因摄像机光学镜头为广角镜头，且光轴与道路平面倾斜，因而采集的道路标线图像会产生鼓形畸变和倾斜畸变。在分析道路标线图像产生鼓形和倾几何斜畸变原因的基础上，采用三次多项式校正道路标线图像的鼓形变形，并提出了一种用几何物像关系校正道路标线图像的倾斜畸变，实验结果表明这种方法是可行的。     

交通事故现场图像拼接方法研究

在交通事故现场勘测中,利用事故现场俯视摄影图,可全面、快速地记录事故现场信息。通过交通事故现场俯视摄影图像几何校正模型对俯视摄影图进行几何校正后,图像存在视场角度小,局部场景丢失的缺点。在分析几何校正坐标变换数学模型的基础上,提出了基于特征线的交通事故现场照片匹配法,实现了图像拼接;并以例证校验了方法的有效性。     

基于时频图像特征约简的柴油机故障特征提取新方法

针对柴油机振动响应信号时频表示特征提取困难的问题,提出了一种直接对柴油机振动谱图像进行特征约简的时频特征提取新方法。通过现有的时频分析方法生成柴油机振动谱图像,采用三次卷积插值法对图像矩阵的维度进行压缩,并将矩阵分块计算的方法与稀疏非负矩阵分解算法(sparse non-negative matrix factorization,SNMF)相结合,用来直接对压缩后的振动谱图像进行特征约简以获取蕴含在其内的低维特征。将提出的方法应用于4种不同状态的柴油机气门故障特征提取试验中,结果证明该方法可准确、快速提取柴油机气门故障特征。     

透视投影空间下的无人机倾斜影像直线匹配

直线匹配在结构化物体的三维轮廓重建方面具有重要的作用,而常规直线匹配方法在处理倾斜影像时效果欠佳。本文从解决航空倾斜影像存在的视角差异问题出发,提出一种透视投影空间下的无人机倾斜影像直线匹配方法。首先,根据地面结构化物体提供的几何约束关系,恢复相机的外方位元素,构建出透视投影矩阵,将原始倾斜影像纠正为近似正形影像,使影像间视角差异减小;然后,对正形影像进行直线提取与匹配,得到匹配线段;最后,利用透视投影矩阵对匹配线段做逆变换,得出原始影像的匹配线段。通过三组无人机航空倾斜影像进行实验可以发现,与直接在原始影像上进行直线匹配相比,此方法在匹配线段的数量和正确率上都有优势。     

基于PCA-SVM的边坡稳定性预测模型研究

提出了一种基于主成分分析(PCA)和支持向量机(SVM)的边坡稳定性预测模型。首先分析了影响边坡稳定性的因素,采用主成分分析方法求取主成分;再将主成分作为输入对支持向量机进行训练,并利用遗传算法优化支持向量机参数;最后通过实例与常用寻参方法所得结果进行比较。结果表明,该法能减少输入变量维数,提高了边坡工程稳定性的预测精度。     

桥梁混凝土结构表面裂缝自动识别技术研究

为了满足桥梁检测发展的需求,提高混凝土结构表面裂缝的识别精度,基于既有图像预处理技术,提出一种新的裂缝提取及宽度计算方法。在图像预处理基础上,通过裂缝连接保留细小裂缝,将裂缝区域当作连通区域,根据裂缝特征参数过滤虚假裂缝,提取真实裂缝;寻找裂缝区域的交叉点,通过交叉点将整个裂缝区域划分成多个小裂缝,将小裂缝的最小外接矩形分割成多个小矩形区域,计算小矩形区域中裂缝特征参数,得到裂缝宽度。以混凝土试件重力加载形成的表面裂缝为对象,改变相机拍摄距离与拍摄倾斜角度获取裂缝图像,应用该方法计算裂缝宽度,并与裂缝观测仪的测量结果进行误差对比分析。结果表明:该算法裂缝识别率高、检测误差小,裂缝检测误差随拍摄距离或拍摄倾斜角度的增大而增大;为了满足实际工程的需要,采集裂缝图像时,拍摄距离最好保持在1m以内,倾斜角度尽量不要超过15°。     

基于LBP算子的路面类型识别研究

不同的路面类型会对车辆的制动、加速、变道等决策产生不同影响,因此实时获取路面类型信息对于提高智能汽车的安全性、舒适性等有着重要意义。论文提出一种基于LBP算子的路面类型识别方法,首先采集了四种车辆行驶常见路面的图像信息,并对图像进行了增广处理;然后使用LBP算子提取出路面图像的纹理特征,再采用PCA方式对纹理特征进行降维;最后通过分类器对数据进行训练与分类。实验结果表明该方法的最高分类准确率可以达到98.5%,有效提升了当前路面类型识别的精度。     

交通事故摄影图像的自标定三维重建及其实现方法

为避免测量过程中在现场设置标定参考点,研究了以基础矩阵估计为基础的交通事故摄影图像自标定模型,并分析了其实现过程。把便于递推实现的无限对称指数滤波器引入平均梯度矩阵计算,实现了图像特征点的检测;通过匹配强度系数计算,对两幅图像的特征点进行了匹配;分析了基础矩阵计算及不良匹配点剔除的最小中值二乘法;对通过本质矩阵分解标定摄影图像的实现过程进行了论述。通过实验对模型及其实现方法进行了验证。     

基于数字图像处理对汽车牌照自动识别系统的研究

以快速准确识别汽车牌照号码为目的,在充分利用数学形态学与多特征组合分析相结合分析的基础上,运用灰度变换、边缘检测、Radon变换、投影特征等图像处理方法,分车牌检测、字符分割、字符识别三步实现汽车牌照的识别,处理过程中考虑并解决了现实拍摄图像中存在的牌照倾斜等不利条件,用MATLAB软件对这些算法进行仿真,经过对多幅图像的处理实验表明,该系统识别速度快,识别率高。     

基于彩色二值化的车牌定位方法

提出了一种基于彩色二值化的车牌定位方法。首先将彩色图像从RGB颜色空间转换到HSI颜色空间,同时生成一个与彩色图像大小相同的二值化状态特征矩阵,根据车牌的色彩特征,调整状态特征矩阵;再用数学形态学方法对状态特征矩阵进行填充空洞和滤除噪声的处理,并根据车牌的几何特征除去伪牌照区。该方法将图像的色彩特征与状态特征分离,充分利用车牌的色彩特征调整彩色图像的状态特征,并融合了数学形态学方法;而且将车牌的色彩特征和几何特征进行了有机的结合。实验结果表明该方法是一种有效的车牌定位方法。     

License plate extraction method for identification of vehicle violations at a railway level crossing

The primary cause of most railroad accidents is vehicle entry into railway level crossings despite warning messages. To identify drivers who violate railway level crossing regulations, vehicle license plate recognition can be applied at railway level crossings. The purpose of this paper is to present an effective method for extracting the license plate region from vehicle images taken at railway level crossings. The method proposed in this paper uses the variation in the gray-level values across the image of a license plate. For license plate region extraction, the character region is first recognized by identifying the character width and the difference between the background region and the character region. The license plate region is then extracted by finding the inter-character distance in the plate region. In addition, the license plate type is identified by the difference in the gray-level value between the background region and the character region. The proposed method is effective in solving the current challenges in extracting the license plate region from the damaged frames of license plates issued for domestic use, including new types of license plates. According to the experimental results, the proposed method yields a high extraction rate of 99.5% for vehicle license plates.     

车牌识别算法的研究与实现

介绍一种基于KL变换的车牌识别算法。算法主要由车牌分割和车牌字符识别组成。车牌分割算法利用车牌的纹理特征和形状特征来定位车牌，并采用投影的方法和先验知识实现车牌的倾斜校正和字符切分；车牌字符识别算法采用一种最优变换，KL变换来提取字符图像的特征值和特征向量，并通过计算待识别字符与各样本字符特征值的欧氏距离来实现对该字符的分类。试验结果表明此方法是正确的。     

Development of a telematics evaluation environment based on a driving simulator and its application

For developing telematics devices, traditional development methods include the unit function test, compatibility test and T-Car, which have some limitations. Telematics devices have various functions that require accounting for the interactions among three major elements of automotive electronics: the vehicle, the device unit and driver. The KAAS (KATECH Advanced Automotive Simulator) system is a virtual-reality-based test environment designed to test and analyze the three elements in one place. One of the difficult functions when constructing such VR (Virtual Reality)-based telematics test environment is to develop a test method for the LBS (Location-Based Service) function such as a car navigation demanding the GPS (Global Positioning System) satellite signals because KAAS is in a fixed laboratory. To overcome these problems, a real-time GPS simulation system, which can be integrated with KAAS, is needed because the location of the vehicle in virtual space is determined purely by the driver’s personal intention while driving virtually. This paper presents new concepts needed to construct a VR-based telematics test environment to generate a GPS RF signal, which reflects the continuously changing vehicle location during virtual driving in real-time. To construct this system, the coordinate transform must be conducted from a rectangular coordinate system that is compatible with a virtual 3D DB that is used to construct a 3D image for KAAS using a WGS84 and a longitude-latitude coordinate system compatible with a GPS simulator. Moreover, the real-time HILS (Hardware In Loop Simulation) systems and the CDMA (Code Division Multiple Access) simulation system are developed to evaluate telematics devices. Finally, we show its applications and results.     

盲图像质量评价在智能交通系统中的应用

针对交通系统图像的特点,构建了盲图像质量评价在智能交通系统中应用的框架。对于获取到的交通图像首先进行预处理（包括车身区域分割和车牌检测）,然后对影响交通图像质量的亮度、对比度及清晰度分别进行评价。对实际获取的图像利用算法进行了验证,实验结果表明,算法可以较好地对交通图像进行分类且评价结果与主观评价一致。     

Dynamic programming-based method for extraction of license plate numbers of speeding vehicles on the highway

In the last decade, vehicle identification systems have become a central element in many applications involving traffic law enforcement and security enhancement, such as locating stolen cars, automatic toll management, and access control to secure areas. As a method of vehicle identification, license plate recognition (LPR) systems play an important role and a number of such techniques have been proposed. In this paper, we describe a method for segmenting the main numeric characters on a license plate by introducing dynamic programming (DP) that optimizes the functionality describing the distribution of the intervals between characters, the alignment of the characters, and the threshold difference used to extract the character blobs. The proposed method functions very rapidly by applying the bottom-up approach of the DP algorithm and also robustly by minimizing the use of environment-dependent image features such as color and edges.     

Slip-Angle Estimation for Vehicle Stability Control

Recently, some direct yaw-moment control systems have been in development. Obviously, such systems need accurate slip-angle information. This paper describes a strategy of vehicle slip angle estimation. The difficulty in slip angle estimation is due to nonlinear characteristics of tyres and influence of relative slant of the road surface. To solve this difficulty, a combined method of model observer and direct integration method is proposed. In this method, two kinds of values of the side forces of the wheels are provided, i.e., direct detected values by the G-sensor and values from a tyre model. Then those values are combined appropriately which results in the combination of model observer and direct integration. A feedback algorithm, redesigned to suppress the influence of tyre model error, is applied in the observer. Considering interference of road surface and its avoidance, road slant angle is estimated and consequently vehicle model is corrected. The estimated value of the road friction coefficient is given by the acceleration, and an adequate bias, depending on yaw-deviation, is added. The calculation method of reference yaw-velocity is improved, in order to avoid interference of road slant and variation of dynamic characteristic of vehicle.     

Slip-Angle Estimation for Vehicle Stability Control

Recently, some direct yaw-moment control systems have been in development. Obviously, such systems need accurate slip-angle information. This paper describes a strategy of vehicle slip angle estimation. The difficulty in slip angle estimation is due to nonlinear characteristics of tyres and influence of relative slant of the road surface. To solve this difficulty, a combined method of model observer and direct integration method is proposed. In this method, two kinds of values of the side forces of the wheels are provided, i.e., direct detected values by the G-sensor and values from a tyre model. Then those values are combined appropriately which results in the combination of model observer and direct integration. A feedback algorithm, redesigned to suppress the influence of tyre model error, is applied in the observer. Considering interference of road surface and its avoidance, road slant angle is estimated and consequently vehicle model is corrected. The estimated value of the road friction coefficient is given by the acceleration, and an adequate bias, depending on yaw-deviation, is added. The calculation method of reference yaw-velocity is improved, in order to avoid interference of road slant and variation of dynamic characteristic of vehicle.     

图像处理技术在机动车检测行业中的应用

数字图像处理技术是当前计算机信息技术应用领域的热门,被广泛地应用到现代机动车检测行业当中,本文主要研究分析了图像处理技术在机动车前照灯检测中的应用,包括采用CCD和CAD技术,分析了机动车前照灯的光学性质和生物学作用。本文还研究了图像处理技术在机动车车牌自动识别系统中的应用、图像处理技术在机动车零件检测中的应用、在轮速检测中的应用以及在机动车类型自动识别当中的作用,充分体现了数字图像处理技术的可靠性和灵活性。     

基于场景分类及灰度跳变的车牌定位方法

车牌定位是自动车牌识别系统实现的关键。提出一种基于场景分类及灰度跳变的车牌定位算法。该算法对彩色图像进行场景分析,将图像分类为白天场景类或夜晚场景类。这两类场景的字符与背景的灰度跳变值不同,一般白天场景类的灰度跳变值较大,夜晚场景类的灰度跳变值较小。利用不同的灰度跳变值快速提取出几块车牌候选区域,对不同的场景用不同的方法最终选取一块区域。实验结果显示本文提出的方法对图像场景分类准确率达到98.2%,车牌定位的准确率达到98.5%。