基于智能手持设备的车牌识别系统研究

以交警能更加快捷方便地稽查运营车辆为目的,借助最新的移动互联网技术,分析并研究了多种计算机视觉算法。提出了一种基于K均值聚类（K-means）的车牌定位技术,通过颜色的聚类分析来确定车牌区域,定位车牌。又应用垂直投影技术和模板匹配法来分割和识别车牌字符,利用SQlite开发了具有存储车牌字符信息功能的数据库。在此基础上,研制了一套基于智能手持设备的车牌识别系统。功能测试表明,该系统具有良好的性能,能够较好地识别出车牌信息。     

基于多核的车牌识别的架构实现

多核技术是现在提高芯片性能的主要方法。区别于传统以PC和DSP为核心的车牌识别系统,以FPGA为核心,利用SOPC技术构建了车牌识别多核处理器。给出了一种基于多核的车牌识别架构,在该多核处理器中,以3个Nios II 软核为主要处理器核处理车牌定位、字符特征识别提取及识别等处理,同时构建硬件加速器作为协处理器处理图像增强、边缘检测和膨胀、腐蚀等数学形态学处理。在CQ片上路由器基础上,构建了NOC用以实现片上多核通信。另外,为了保证路由器与多处理器核之间的快速、并行通信,加入了数据驱动模块。整个系统在Altera Cyclone IV FPGA上实现了车牌的识别。这种片上系统设计方法具有硬件设计灵活,可扩展性强等优点,能有效地降低系统软硬件设计的难度,缩短开发周期,并提高设计的可靠性。     

基于LabVIEW图像法车牌智能识别系统

车辆牌照自动识别是实现交通管理智能化的重要环节,设计中利用图像采集卡对经过的车辆车牌进行图像采集并传送至计算机,采用美国NI公司LabVIEW软件,实现图像预处理、图像去噪以及图像增强等功能;然后根据车牌颜色特征对其准确定位,采用阈值法分割车牌字符;最后由OCR函数来识别字符,识别结果保存至相应数据中,可以进行相应的违章、违规智能交通管理,经实验该系统成功实现车牌识别识别率达99%。     

基于BP神经网络的车牌字符识别算法研究

车牌字符识别是车牌识别系统中的关键环节。采用图像处理和神经网络相结合的方法设计新的车牌字符识别算法,先对分割出的车牌字符进行归一化处理,然后进行SOBEI．边缘检测和角点特征提取,最后输入BP神经网络进行训练、识别,其中BP神经网络模型属于改进型神经网络。通过一系列神经网络训练和仿真实验,车牌识别速度和正确率得到了明显的提高。     

基于区域对比度和SSIM的图像质量评价方法

在诸多图像质量评价方法中,结构相似度(SSIM)算法简单高效,准确性较高,但SSIM模型不能很好地评价存在局部失真和交叉失真类型的图像。针对SSIM算法对图像不同区域平等对待的不足并考虑了时域人眼视觉特性,提出一种改进的基于区域对比度和结构相似度(RCSSIM)的图像质量评价方法。该算法将图像区域灰度信息对比度与SSIM算法融合,加权归一为参考图像与失真图像的对比度结构相似度值,以其评价图像质量。在LIVE图像数据库上的实验结果表明,与SSIM算法相比,RCSSIM评价结果的皮尔逊线性相关系数提高约0.015,均方根误差减小约0.55,更接近于人眼主观测试结果,具有更好的评价性能。     

一种新型自适应环境光照的车牌定位算法

为解决在不同光照环境下车牌的适应性定位问题,同时解决基于颜色定位的方法普遍存在的车身颜色与车牌颜色相近,以至于难以区分的问题,分析了不同光照下的车牌图像在HSV空间的分布特性与图像亮度等参数关系,结合车牌的混色分布特点、RGB与HSV空间变量转换关系,提出了一种可直接对RGB空间进行处理、能适应光照环境的快速车牌定位算法;实验结果表明,提出的定位算法对光照不足、环境光不均匀等复杂条件下获取的蓝白车牌图像具有很高的定位可靠性。     

基于灰度跳变与字符间隔模式的车牌定位方法研究

车牌定位是车牌自动识别的第一步,而如何考虑光照影响是车牌定位是否成功的关键;通过深入分析不同的车牌图像,提出一种基于灰度跳变与字符间隔模式的车牌定位方法;首先,针对不同光照条件下采集到的车牌图像明暗度的不同,利用多阈值处理方法得到车牌信息不丢失的、最佳的二值图像,然后,在其二值图像中首先利用灰度跳变定位车牌的上下边界,接着对字符垂直投影后的宽度进行统一的调整并以固定的字符间隔特征定位车牌的左右边界,从而完成车牌定位;最后,通过实验验证了该方法的有效性。     

基于自适应耦合PDE模型的车牌图像去噪研究

针对车牌识别预处理中的图像去噪问题,提出一种自适应耦合偏微分方程(PDE)去噪模型;该模型在各项异性扩散模型的基础上,构造一种新的去除椒盐噪声的扩散项,能够根据噪声图像特点自适应控制扩散速度,有效抑制椒盐噪声,并将新的扩散项与各向异性扩散模型进行耦合,并提出一种新的耦合系数计算方法,根据图像信息自适应计算耦合系数,使得新模型能够在新的扩散项和各项异性扩散模型间自适应转换,有效去除车牌图像中的混合噪声;为了抑制去噪引起的图像边缘模糊问题,引入振动滤波进行逆滤波,增强图像的边缘信息;实验结果表明,自适应耦合PDE模型能更有效去除车牌图像中的混合噪声,保护图像的边缘信息,提高图像的峰值信噪比(PSNR);去噪后的图像更有利于后续的字符分割与识别,有效提高车牌图像的识别准确率。     

基于投影寻踪的高光谱典型目标识别算法

针对典型目标识别问题,提出了一种基于投影寻踪的高光谱目标识别算法。先对高光谱图像进行最小噪声分离变换,计算出本征维度,同时对图像去噪,然后采用信息散度作为投影指标,对投影指标值自适应分割,得到所要提取的波谱曲线,最后用光谱角匹配识别出目标及其位置。高光谱图像验证结果表明,该方法有效地去除了图像噪声,而且能够快速、可靠地提取端元并识别出目标。     

未系安全带智能识别中安全带定位关键方法研究

安全带定位是实现机动车未系安全带智能识别的关键。针对道路监控图像特点,提出一种基于梯度变换的安全带定位方法。该方法在对卡口图像进行预处理的基础上,采用自适应阈值边缘检测算法以及积分投影方式定位车辆位置,设立车牌检测区域,以减少运算量,降低干扰,同时利用训练得到的Haar分类器识别车牌位置,通过逼近方式切取车辆右侧图像。最后采用梯度变换算法求得车窗各边缘坐标,实现车窗精准定位,并计算得到安全带位置。试验表明,该方法可实现安全带的准确定位,具有较好的实用性,为后续安全带识别奠定基础。     

License plate location based on trichromatic imaging and color-discrete characteristic

License plate location (LPL) is the key part of automatic license plate recognition (LPR) system that plays an important role in many applications and a number of LPL methods based on color information have been proposed. However, some problems, such as country specific, similar color interference, lighting variation sensitive, time consuming, processing only one kind of LP with same color combination a time and omitting the case of more than one combinations in a LP, should be solved further. In this study, a color-based LPL method that consists of three modules: color edge extraction, denoising and searching, is proposed. Color edge extraction, the kernel of the proposed, is designed by color-discrete characteristics of license plates in the trichromatic wavebands. In the experiment, 1384 images taken from natural scenes in China and other 104 countries or regions are employed. Of which, 74 have been failed to locate the license plates. The success rate and average execution time are 94.7% and 57 ms, respectively.     

Multi-scale SSIM metric based on weighted wavelet decomposition

Image quality assessment aims to use computational models to assess the image quality consistently with subjective evaluations. This paper proposes a new metric composed of weighted wavelet multi-scale structural similarity (WWMS-SSIM). Four-level 2-D wavelet decomposition is performed for the original and disturbed images, respectively. Each image can be partitioned into one low-frequency subband (LL) and a series of octave high-pass subbands (HL, LH and HH). Different subbands are processed with different weighting factors. Based on the results of the above, we can construct a modified WWMS-SSIM. Comparison experiments show that the correlation, prediction accuracy and consistency of the proposed metric are respectively 5.8%, 5.2% and 4.8% higher than the PSNR metric. The correlation, prediction accuracy and consistency of the proposed metric are respectively 0.7%, 1.6% and 2.6% higher than the SSIM metric. In terms of the experiment results, the WWMS-SSIM metric shows good feasibility comparing with PSNR and SSIM methods.     

基于对称特征相似性的图像识别算法研究

研究了一类微小生物——线虫的图像识别算法。首先对图像进行边缘检测、二值化和去除干扰点等操作,得到线虫的身体轮廓线;然后利用其身体的对称特征,通过比较相邻线段的相似性实现对线虫的识别及对干扰的区分。实验结果表明,此算法优于骨架法、面积法等典型图像识别方法,对线虫的检测和定位具有很高的准确度和快速性,并可应用于其它具有对称特性的微小生物的图像识别。     

基于K-L变换的虹膜识别方法

提出了一种基于K-L变换和最近邻分类器的虹膜识别方法。该方法采用K-L变换得到一组虹膜图像基,并利用这组基图像构造一子空间,将待识别图像在这个空间上的投影系数作为待识别图像的特征向量。采用最近邻分类器进行了分类。基于CASIA虹膜数据库的试验表明,将K-L变换应用在虹膜特征提取上可以取得较高的识别率。     

Image quality assessment in multiband DCT domain based on SSIM

The performance of image quality assessment method based on SSIM (structural similarity) is better than the PSNR (peak signal to noise ratio), but the assessment effects of SSIM is poor for seriously blurred image, therefore, the model that combined HVS (human visual sensitivity) and SSIM was established. The basic idea is based on the human eye's sensitivity to different frequency distortion image, the image is two-dimensional discrete cosine transform frequency component into low, mid, high-frequency component, to obtain the frequency component of light, contrast and structural information, using Pearson coefficient for weight and sum processing to the sub-image according to frequency bands of different sensitive degree, finally, get the sharpness of the image. Through nonlinear regression analysis of objective assessment and DMOS, experiments showed that this method was closer to human perception than SSIM and GSSIM for serious blurred distortion image. At the same time, compared to conventional algorithm MAE (mean absolute error), MSE (mean square error) and PSNR, this model was more consistent with human visual characteristics.     

A face recognition algorithm based on collaborative representation

In this paper, we propose a face recognition algorithm by incorporating a neighbor matrix into the objective function of sparse coding. We first calculate the neighbor matrix between the test sample and each training sample by using the revised reconstruction error of each class. Specifically, the revised reconstruction error (RRE) of each class is the division of the l₂-norm of reconstruction error to the l₂-norm of reconstruction coefficients, which can be used to increase the discrimination information for classification. Then we use the neighbor matrix and all the training samples to linearly represent the test sample. Thus, our algorithm can preserve locality and similarity information of sparse coding. The experimental results show that our algorithm achieves better performance than four previous algorithms on three face databases.     

基于小波不变矩的多类目标特征选择算法

特征选择是目标识别中的重要问题。对于有着3个参数（m,n,q）的小波矩来说更是如此。基于2类模式的特征选择思想,提出多类模式下绝对可分的特征选择算法,给出图像数值化处理中参数(m,n,q)的合理取值范围。实验结果表明：无论是对差别比较大、差别比较小还是混合型的多类目标,经过此特征提取出来的小波矩都有着较好的识别效果。     

Finger vein recognition algorithm based on optimized GHT

In this paper, we present a novel algorithm for recognition of vein features based on optimized generalized Hough transform (GHT). The new algorithm involves several steps. First, it extracts singular points from the binary image of finger veins, and segments the finger veins by these points. Then it selects valid segments and sequences them by way of chain codes. Next it uses the optimized GHT to differentiate sectional curves of finger veins from the whole finger vein image. Using this approach reduces the influence of fragmentation, enhances adaptability for displacement, rotation, and zooming, and accordingly improves the quality of finger vein recognition. We have tested the proposed method with actual finger vein images and produced very satisfactory reassembly results.     

基于深度学习的ELM实时识别研究

基于深度学习的方法,在HL-2A装置上开发出了一套边缘局域模(ELM)实时识别算法。算法使用5200次放电数据(约24.19万数据切片)进行学习,得到一个深度为22层的卷积神经网络。为衡量算法的识别能力,识别了HL-2A装置自2009年实现稳定ELMyH模放电以来所有历史数据(约26000次放电数据),共识别出1665次H模放电,其中误识别35次,误报率为2.10%。在实际的1634次H模放电中,漏识别4次,漏识别率为0.24%。该误报率和漏报率可以满足ELM实时识别的精度要求。识别算法在实时控制环境下,对单个时间点的平均计算时间为0.46ms,可以满足实时控制的计算速度要求。     

基于改进HOG特征提取的车型识别算法

本文针对高速环境下的车型识别问题,提出基于方向可控滤波器的改进HOG算法。将方向可控滤波器算法与HOG算法相结合,以实现对车辆图像特征提取。采用主成分分析算法(PCA)约减特征向量维数以减少计算复杂度,利用支持向量机算法对提取特征进行样本训练,实现对车辆外型特征的识别。仿真实验结果表明:采用该算法原始车辆车型的识别正确率均值达到92.36%;另外,本文方法的识别速度比传统的HOG特征算法提高了3.45%,识别实时性得到提升。本文算法比传统HOG算法更优,能有效提高车型识别的效率。