基于颜色搭配与纹理特征的车牌定位方法

牌照定位是牌照识别系统中的关键技术。目前在多数牌照定位方法中考虑到了牌照的颜色和纹理特征,但对不同环境的适应性不强。为解决这方面的问题,首先从机动车牌照具有固定颜色搭配的特点出发构造颜色搭配掩模矩阵,并利用此掩模矩阵对原边缘检测图像进行条件约束,得到约束二值边缘图像;然后应用具有去噪能力的形态学结构元,形成牌照粗定位候选区域;最后依据牌照的纹理特征从候选区域中提取出真正的牌照。采用了BP神经网络获得强适应性的HSI空间牌照颜色识别方法,并且只在边缘点邻域内实现颜色空间转换运算,能极大地缩减定位周期。经实验表明,该方法能在复杂的环境和不同光照条件下快速地实现不同牌照的精确定位。     

一种梯度特征与区域合并的车牌定位方法

车牌定位是车牌识别系统的重要组成部分。针对车牌目标所在区域梯度变换频繁的特点,利用高通滤波保留梯度变换频繁区域,形态学处理后,将相邻区域进行合并以确定车牌的候选区域。再结合车牌的几何特征与区域目标背景比,找到车牌位置,利用投影方法去除车牌边框,实现车牌的精确定位。实验结果表明,该方法削弱了传统车牌定位算法对车辆大小、图像环境、拍摄角度等的要求,进一步提高了算法的鲁棒性和实用性。     

自然场景下变形车牌检测模型DLPD-Net

目的 随着智能交通领域车牌应用需求的升级,以及车牌图像复杂性的提高,自然场景下的车牌识别面临挑战。为应对自然场景下车牌的不规则变形问题,充分考虑车牌的形状特征,提出了一种自然场景下的变形车牌检测模型DLPD-Net （distorted license plate detection network）。方法 该模型首次将免锚框目标检测方法应用于车牌检测任务中,不再使用锚框获取车牌候选区域,而是基于车牌热力值图与偏移值图来预测车牌中心;然后基于仿射变换寻找车牌角点位置,将变形车牌校正为接近于正面视角的平面矩形,从而实现在各种自然场景下变形车牌的检测。结果 一方面,基于数据集CD-HARD评估DLPD-Net检测算法的性能;另一方面,基于数据集AOLP （the application-oriented license plate database）和CD-HARD评估基于DLPD-Net的车牌识别系统的有效性。实验结果表明,DLPD-Net具有更好的变形车牌检测性能,能够提升车牌识别系统的识别准确率,在数据集CD-HARD上识别准确率为79.4%,高出其他方法4.4% 12.1%,平均处理时间为237 ms。在数据集AOLP上取得了96.6%的识别准确率,未使用扩充数据集的情况下识别准确率达到了94.9%,高出其他方法1.6% 25.2%,平均处理时间为185 ms。结论 本文提出的自然场景下的变形车牌检测模型DLPD-Net,能够实现在多种变形条件下的车牌检测,鲁棒性强,对遮挡、污垢和图像模糊等复杂自然环境下的车牌检测具有良好检测效果,同时,基于该检测模型的车牌识别系统在非受限的自然场景下具有更高的实用性。     

基于混合特征的车牌定位算法

车牌定位技术是汽车牌照自动识别和智能交通系统的用车牌的颜色、纹理和结构几何等多维特征,实现车牌定位.该算法利用车牌的彩色信息进行彩色分割,实现车牌图像的二值化,而后提取边缘增强,在此基础上利用数学形态学方法去噪并去除车牌边框,并利用车牌纹理特征利用投影实现车牌的最终定位.该算法克服了单一特征信息不完备引起的车牌定位误差,实验表明该方法具有较好的车牌定位效果.     

基于扩展数学形态学的车牌定位算法

在汽车牌照识别系统中,车牌定位是整个识别模块实现的前提,目前车牌定位的方法多种多样,各有所长,但存在着计算量大或定位准确率不高等问题。边缘检测是常用的车牌定位方法,边缘检测的质量决定了车牌图像的最终定位结果。一般人们习惯于用基于梯度和基于模板的算子提取边缘,但这类算子都不能很好地滤除噪声,因而给噪声图像边缘检测带来了困难。根据数学形态学原理与方法,提出一种扩展数学形态学车牌图像边缘检测算子,并结合水平和垂直投影进行车牌定位。实验结果表明,该算法不仅能成功提取车牌图像边缘,而且能很好地滤除噪声,从而实现准确车牌定位。     

基于模糊边缘检测算法的车牌定位

车牌定位是自动车牌识别系统的一个关键步骤,车牌定位结果直接影响对车牌的最终识别效果。因此为了保证实际应用中车牌的识别准确率,文中提出了一种新的车牌定位算法,该算法利用一种改进的快速模糊边缘检测算法来进行车牌图像的边缘检测,得到整个原车牌图像的边缘图像,然后基于边缘图像的丰富的边缘信息设计一个高效、准确的车牌区域定位算法,检测出车牌区域。实验结果表明:算法定位速度较快、准确度较高,具有良好的应用前景。     

基于形态学梯度重建的车牌定位方法

准确定位车牌是车牌识别的重要基础。针对复杂环境下车牌图像容易受背景、光照等因素的影响而导致车牌定位精度较低的问题,提出了一种基于形态学梯度重建的车牌定位方法。该方法首先利用颜色信息确定车牌候选区域;然后利用矢量梯度算子获取候选区域中车牌图像的梯度,利用形态学梯度重建运算提取具有车牌特征的图像结构,同时抑制非车牌特征的图像结构;最后利用车牌固定的宽长比先验信息对矩形区域进行提取,最终得到准确的车牌定位结果。实验结果表明,提出的车牌定位方法能在复杂环境下快速、准确地定位车牌,且具有较高的鲁棒性和实时性。     

汽车牌照识别技术研究

本文分析了汽车牌照的几何特征和成像特点,提出了一种基于边缘检测和Hough变换的汽车牌照定位方法和基于图像投影的车牌字符分割方法。通过分析车牌号码中的字符图像特点,提出了基于字符图像几何形态和笔画结构的字母和数字识别方法,以及基于汉字结构知识的汉字识别方法,从而实现了汽车牌照的自动识别。     

自适应不同条件的车牌检测算法

车牌检测(License Plate Detection, LPD)是自动车牌识别中(Automatic License Plate Recognition, ALPR)主要步骤,针对不同条件下车牌检测速度慢和检测精度低的问题,提出了一种改进改进自适应形态闭和开操作的车牌检测算法,该算法首先采用局部直方图对车牌图像均衡化处理,使用自适应形态闭操作对所有灰度化区域进行平滑处理,之后引入局部自适应阈值处理,能得到平滑图像和被分离的车牌,最后采用形态学开操作,将外部区域和车牌数连接部分分离。实验结果表明,所提方法的检测精度高于其他算法,同时,平均检测时间少于其他算法,适合不同条件下实时车牌检测。     

脉冲耦合神经网络在车牌定位中的应用

针对冗余边缘对基于边缘统计特征的车牌定位算法存在较严重干扰的问题,提出一种基于脉冲耦合神经网络(PCNN)的车牌定位方法。在借鉴传统算法的基础上,为抑制干扰性边缘,引入简化的PCNN模型,仅对候选区进行数次PCNN迭代运算,可大幅降低运算复杂度并提高车牌定位率。对300幅车辆图像进行仿真实验,取得了98.3%的定位率。     

Real-time automatic license plate recognition for CCTV forensic applications

We propose an efficient real-time automatic license plate recognition (ALPR) framework, particularly designed to work on CCTV video footage obtained from cameras that are not dedicated to the use in ALPR. At present, in license plate detection, tracking and recognition are reasonably well-tackled problems with many successful commercial solutions being available. However, the existing ALPR algorithms are based on the assumption that the input video will be obtained via a dedicated, high-resolution, high-speed camera and is/or supported by a controlled capture environment, with appropriate camera height, focus, exposure/shutter speed and lighting settings. However, typical video forensic applications may require searching for a vehicle having a particular number plate on noisy CCTV video footage obtained via non-dedicated, medium-to-low resolution cameras, working under poor illumination conditions. ALPR in such video content faces severe challenges in license plate localization, tracking and recognition stages. This paper proposes a novel approach for efficient localization of license plates in video sequence and the use of a revised version of an existing technique for tracking and recognition. A special feature of the proposed approach is that it is intelligent enough to automatically adjust for varying camera distances and diverse lighting conditions, a requirement for a video forensic tool that may operate on videos obtained by a diverse set of unspecified, distributed CCTV cameras.     

复杂背景下的混合特征车牌定位方法

针对各种复杂背景的车牌定位问题,提出一种复杂背景下基于车牌混合特征的车牌定位算法。首先对彩色图像进行预处理,并利用基于边缘检测方法进行二值化;然后结合横向数学形态学运算和车牌几何形状特征,提取出矩形车牌候选区域;最后根据车牌颜色特征在HIS空间下结合垂直和水平投影对车牌区域进行精确定位。实验表明,该算法适用于任意大小、位置和背景环境下的车牌定位,能有效解决仅仅依靠纹理信息或颜色信息车牌定位率低的问题,具有较强的鲁棒性。     

Design and research of automobile anti-collision warning system based on monocular vision sensor with license plate cooperative target

At present, detection method for the target vehicle based on monocular vision sensor uses the whole vehicle as targets. The automobile anti-collision technology proposed in this paper adopts monocular vision sensor for automobile measurement based on vehicle license plate cooperative target. Monocular vision sensor has advantages of good real-time performance and low cost. The technique can improve the detection capability of vehicle collision avoidance warning systems. In addition to the target vehicle positioning, it can also realize attitude determination. This technology eliminates the limits of road surface roughness and fluctuation. This paper designs the realization scheme of collision warning system based on monocular vision sensor from the automobile license plate cooperative target. Technology roadmap of automobile collision warning system is given. In this paper, license plate frame location is as the research background. The paper presents an analytic solution of positioning method for the license plate frame image. The method uses four vertex characteristics of license plate frame image to locate. Positioning speed of the method is fast. And it has a unique solution. This method can be used to positioning for the license plate frame. Simulation experiment is done for the collision warning location. The simulation results show that this method can locate the position for license plate frame image. License plate is regular shape, uniform, with identity recognition function markers on the automobile body. In the previous research on automotive collision warning and intelligent vehicle, we have not seen the research methods similar to the method introduced in this paper. The research enriches automobile anti-collision technology and theory of intelligent vehicle technology. It can also provide an auxiliary method for navigation and obstacle avoidance research for unmanned vehicle. It has certain scientific significance. Vehicle collision warning system can help the driver judgment, prompting warning, improving driving safety, and has broad application prospects.     

基于MATLAB的车牌定位算法的分析与仿真

车辆牌照识别是智能交通系统的重要部分。该论文针对车辆牌照识别问题中的车牌定位问题进行了分析,并对车牌图像的灰度处理,车牌图像的边缘检测,车牌图像的形态学处理等相结合的方法对车辆牌照进行仿真分析。实践证明,与其他的车牌区域定位方法相比,用该方法要相对快一些,而且成功率高一些。     

基于主成分分析网络的车牌检测方法

车牌识别是智能交通系统的核心技术,车牌检测是车牌识别技术中至关重要的一步。传统的车牌检测方法多利用浅层的人工特征,在复杂场景下的车牌检测率不高。基于主成分分析网络的车牌检测算法,能够无监督地逐级提取车牌深层特征,可有效提高算法的鲁棒性。算法首先采用Sobel算子边缘检测和边缘对称性分析获取车牌候选区域;然后将候选区域输入到主成分分析网络中进行车牌深度特征提取,并利用支持向量机实现对车牌区域的判别;最后采用非极大值抑制算法标记最佳车牌检测区域。利用收集的复杂场景下的车辆图像对所提方法的参数进行分析,并将其与传统方法进行比较。实验结果表明,所提算法的鲁棒性高,性能优于传统的车牌检测方法。     

Automatic License Plate Recognition via sliding-window darknet-YOLO deep learning

Automatic License Plate Recognition (ALPR) is an important research topic in the intelligent transportation system and image recognition fields. In this work, we address the problem of car license plate detection using a You Only Look Once (YOLO)-darknet deep learning framework. In this paper, we use YOLO's 7 convolutional layers to detect a single class. The detection method is a sliding-window process. The object is to recognize Taiwan's car license plates. We use an AOLP dataset which contained 6 digit car license plates. The sliding window detects each digit of the license plate, and each window is then detected by a single YOLO framework. The system achieves approximately 98.22% accuracy on license plate detection and 78% accuracy on license plate recognition. The system executes a single detection recognition phase, which needs around 800 ms to 1 s for each input image. The system is also tested with different condition complexities, such as rainy background, darkness and dimness, and different hues and saturation of images.     

Statistical characteristics in HSI color model and position histogram based vehicle license plate detection

Visual perception takes an important role in the implementation of intelligent robot and transportation systems. Such perception is to detect and recognize various objects in the real environment. Detecting license plate (LP) is a crucial and inevitable component of the vehicle license plate recognition (VLPR) system. In this proposed algorithm, initially, HSI color model is adopted to select automatically statistical threshold value for detecting candidate regions. According to different colored LP, these candidate regions may include LP regions; geometrical properties of LP are then used for classification. The proposed method is able to deal with candidate regions under independent orientation and scale of the plate. Finally, the decomposition of candidate regions contains predetermined LP alphanumeric character by using position histogram to verify and detect vehicle license plate (VLP) region. In experiment more than 150 images were used, and they were taken from the variety of conditions such as complex scenes, illumination changing, distances and varied weather etc. Under these conditions, success of LP detection has reached more than 94%.     

车辆牌照定位算法研究

车牌定位是车牌自动识别系统中的一个关键问题。提出了一种简单高效的车牌定位算法。在分析车牌图像的特征后,先利用一系列图像处理,然后进行模糊模板匹配,最后对匹配到的车辆牌照候选区分别加以验证,即得到确切的车辆牌照子图像区域,大量实验数据和现场测试证明,车辆牌照图像定位准确率达96％,取得了很好的系统性能和实效。     

A novel license plate location method based on wavelet transform and EMD analysis

Although various license plate location methods have been proposed in the past decades, their accuracy and ability to deal with different types of license plates still need to be improved. A robust license plate location method can raise the accuracy of the whole license plate recognition procedure. This paper proposes a robust method based on wavelet transform and empirical mode decomposition (EMD) analysis to search for the location of a license plate in an image to deal with some challenging problems in practice such as illumination changes, complex background and perspective change. By applying wavelet transform on a vehicle image and projecting the acquired details of the image, a wave crest that indicates the license plate will be generated. In order to locate the desired wave crest in the nonlinear and non-stationary projection dataset, EMD analysis is applied. Using the reconstructed projection data and the Hilbert transform of intrinsic mode function components, the position of the license plate is detected. Comprehensive experiments show that this method can locate the positions of various types of license plates with a high accuracy of 97.91% and a relatively short running time.     

A new and robust method for character segmentation and recognition in license plate images

This paper provides a new and fast method for segmentation and recognition of characters in license plate images. For this purpose, various methods have been proposed in literature. However, most of them suffer from: sensitivity to non-uniform illumination distribution, existence of shade in license plate, license plate color and the need for receiving an exact image of the license plate. In the proposed algorithm, non-uniform illumination and noise are reduced by a Gaussian lowpass filter and also by an innovational Laplacian-like transform and characters are segmented by a set of indigenous and relative features. To be prepared for recognition, the segmented characters are normalized by a local algorithm. Two feed-forward neural networks with back-propagation learning method are employed for character recognition. The principal component analysis (PCA) is used to decrease input data and, consequently, computational complexity. The proposed algorithm does not necessarily need an exact plate image and can receive a band from the vehicle original image as an input, which includes the plate. Our proposed method is completely robust to the disturbances such as non-uniform brightness distribution on the various positions of a license plate image and the plate color. In order to evaluate our algorithm, we applied it on a database including 120 vehicle images with different backgrounds, plate colors, brightness distributions, distances and viewing angles. The results confirm the robustness of the proposed method against severe imaging conditions.