Orthogonal discriminant improved local tangent space alignment based feature fusion for face recognition

Improved local tangent space alignment (ILTSA) is a recent nonlinear dimensionality reduction method which can efficiently recover the geometrical structure of sparse or non-uniformly distributed data manifold. In this paper, based on combination of modified maximum margin criterion and ILTSA, a novel feature extraction method named orthogonal discriminant improved local tangent space alignment (ODILTSA) is proposed. ODILTSA can preserve local geometry structure and maximize the margin between different classes simultaneously. Based on ODILTSA, a novel face recognition method which combines augmented complex wavelet features and original image features is developed. Experimental results on Yale, AR and PIE face databases demonstrate the effectiveness of ODILTSA and the feature fusion method.     

Traffic conflict assessment for non-lane-based movements of motorcycles under congested conditions

Traffic conflict under congested conditions is one of the main safety issues of motorcycle traffic in developing countries. Unlike cars, motorcycles often display non-lane-based movements such as swerving or oblique following of a lead vehicle when traffic becomes congested. Very few studies have quantitatively evaluated the effects of such non-lane-based movements on traffic conflict. Therefore, in this study we aim to develop an integrated model to assess the traffic conflict of motorcycles under congested conditions. The proposed model includes a concept of safety space to describe the non-lane-based movements unique to motorcycles, new features developed for traffic conflict assessment such as parameters of acceleration and deceleration, and the conditions for choosing a lead vehicle. Calibration data were extracted from video clips taken at two road segments in Ho Chi Minh City. A simulation based on the model was developed to verify the dynamic non-lane-based movements of motorcycles. Subsequently, the assessment of traffic conflict was validated by calculating the probability of sudden braking at each time interval according to the change in the density of motorcycle flow. Our findings underscore the fact that higher flow density may lead to conflicts associated with a greater probability of sudden breaking. Three types of motorcycle traffic conflicts were confirmed, and the proportions of each type were calculated and discussed.     

关注智能汽车安全问题刻不容缓

<正>"高速上,一辆疾驰的汽车突然失去控制,刹车失灵,车主在大多数情况下因此而丧生,镜头的另一端是谋杀者或狡黠或冷漠的眼神。"这是电影里的常见镜头。我们多数情况下会想到是刹车被做了手脚。然而,现在的技术告诉我们,汽车通过网络可以被随意控制绝不是一个传说。当越来越多品牌把车联网应用当做一种时髦,为外部网络访问提供更多应用软件平台和越来越多的远程信息处理接口时,每一辆车在黑客眼里     

路面裂缝图像自动识别算法综述

路面裂缝自动检测对于路面养护管理、路面性能评价与预测、路面材料和结构设计具有重要的实用价值,但快速、准确、全面且稳定地识别路面裂缝一直是个难题.为此,对路面裂缝自动检测研究现状进行综述,包括以图像增强和去噪为目的的预处理方法,基于阈值分割、边缘检测和种子生长的空间域识别算法,以小波变换为代表的频域识别算法,基于有监督学习的识别算法及其他裂缝识别方法;指出既有裂缝识别算法存在易受光照和油污等因素的影响、裂缝识别图像连续性差和识别速度和精度较低等不足.最后,提出综合考虑边界和区域特征消除纹理和噪声干扰、基于局部和全局信息设计优化识别算法和基于三维图像进行裂缝识别等研究展望,为裂缝自动识别算法的改进提供参考.     

基于视频图像处理的微观交通数据的获取

微观交通数据对于通行能力的分析、智能交通系统的建立、异常交通状态的识别、微观交通流特性的研究以及微观交通仿真技术的发展具有重要意义.随着视频采集技术在交通领域应用的日益增加以及图像处理技术的发展,使得获取微观交通数据成为可能.鉴于此,利用摄影学原理,实现了图像坐标与实际坐标的转换,并从视频中提取了车辆的行驶轨迹、车速和车头时距,为研究微观交通系统提供了数据支持.     

高速公路二义性路径识别系统设计

随着路网规模和交通量的增大,联网收费建设进程日益加快,用户对高速公路通行效率和服务水平的要求也逐渐增高.由于道路投资主体增多,路网结构变得更为复杂,联网收费范围扩大以后,因为二义性路径的存在,道路投资者之间产生的矛盾越来越突出.鉴于此,通过利用目前先进的远距离RFID（射频识别）技术,对高速公路二义性路径识别管理系统进行了分析和研究.该技术能够实现对路径的精确识别,解决高速公路业主方利益分配问题.     

测试系统中波形的神经网络识别方法

为满足自动测试系统对信号波形的智能化识别要求,在对现有的波形识别方法进行研究的基础上,提出一种新的波形识别方法;该方法在利用离散余弦变换（DCT）对波形进行频域特征提取的基础上,将模式识别与BP神经网络理论相结合从而建立波形特征匹配模板,最终实现对信号波形的智能识别;对比试验表明,该方法与最大相关系数法（MCC）、最大相关差值法（MCD）相比很好地解决了待测波形与模板波形匹配过程中的相位对齐过程,而且更加快速、高效。     

基于综合特征的字符模板库的建立与训练

叙述邮政编码自动分拣系统的软件处理的原理以及程序设计方法,对于倾斜放置信封的邮编区域的定位问题提出解决方案,利用信息熵理论对字符的综合特征的选取提出融合方案,建立训练了邮政编码识别系统中的手写数字模板库,并设计与之相关的可视数据库系统,该系统大大提高信封的自动分拣率以及可行性。     

移动模架法施工连续梁桥线形控制探讨

线形控制是采用移动模架施工连续粱桥的关键技术之一,根据桥面线形控制的原理和方法,结合移动模架法的施工特点,提出了凫洲大桥移动模架法现浇连续箱梁施工的内外模平面线形控制、高程控制、线形监测及调整方案,工程实践表明本文的方案可行。可以为类似工程施工线形控制提供参考。     

铁路路况识别系统的设计和开发

为了实现对多种复杂铁路路况的识别,需要采集路况图像并加以处理,以得到识别所需的图像特征,最后通过模式识别辨别路况.由摄像机、采集卡、计算机及相关软件建立一套路况图像采集与识别系统.探讨数字图像的采集、图像处理中同态滤波、对比度增强、优化二值化算法、膨胀操作、粒子滤波等处理方法以及模式识别方法在铁路路况识别系统中的应用.经过现场验证,所设计系统识别准确率达94.4%,识别效果理想,可以满足路况采集及识别的需要,是一种可以运用在铁路机车智能减振系统中的新方法.