畸变汽车牌照图像的空间校正

首先针对我国汽车版照的常规模式，提出了用Hough变换提取牌照边框参数的方法，进而求得汽车牌照四边形区域4个顶点的坐标，并且给出了利用对极坐标参数表示的直线进行Hough变换的具体途径。然后，根据畸变牌照的4个顶点和归一化汽车牌照图像的4个顶点之间的一一对应关系，求解双线性空间变换方程，进而对畸变图像进行校正。最后给出了一种图像毛刺的剔除方法。实验表明，本文介绍了的方法对畸变牌照图像的归一化可以取得良好的效果。     

成像偏振在车道线检测与识别中的应用

复杂环境下的车道检测是目前智能车和辅助安全驾驶研究的难点和热点. 针对外部复杂的道路环境,将光学偏振理论引入传统的车道检测技术,提出了一种基于成像偏振的车道线检测方法. 通过对车道线图像基本特征的分析,首先采集3个角度的特殊环境道路偏振图像,获得偏振度图像;然后对偏振度图像作二值化和图像感兴趣区域的划分;再根据车道线边缘的直线特性,进行道路图像的边缘检测从而可以获得车道边缘;最后通过Hough变换原理提出了改进的Hough算法,并得以实现检测出车道标线,计算出汽车行驶偏角. 通过仿真和实验验证表明,该方法能够准确地检测和识别出复杂环境下的车道线,车道线的检测偏角与实际偏角之间的误差小于0.3°.      

梯度Hough变换在圆检测中的应用

介绍了一种有效的利用图像梯度信息的Hough变换方法——梯度Hough变换(GHT).详细讨论了梯度的求取、GHT的详细算法,讨论了GHT算法和直接累加Hough变换算法的区别及改进之处.     

一种基于形态学变换的车道检测方法

利用形态学图像处理技术，基于形态学变换的道路检测方法，能够应用形态学变换、Canny边缘检测与Hough变换检测出直线道路车行道的边缘线。试验表明该方法能够准确提取出目标区域轮廓。     

人脸识别系统中的特征提取

研究了人脸识别系统中正面人脸的特征提取,通过区域增长从人脸图像中分割出人脸,再利用边缘检测,Hough变换,模板匹配和方差投影技术可以快速有效地提取出人脸面部器官眼睛,鼻子和嘴巴特征,实验结果表明本文所采用的方法具有较高的准确率和光照鲁棒性。     

Hough变换在汽车纵梁漏孔检测中的应用

结合纵梁漏孔检测的具体情况对随机Hough变换加以改进和应用,实现对纵梁图像圆的检测,并减少了随机Hough变换中存在的无效积累问题。     

图像处理对影像法检测改进算法的研究

在数字万能工具显微镜上进行影像法测量,提出一种基于CCD图像技术的万工显改进方法.针对工件影像的特点,提出一种基于灰度变换,Sobel边缘检测和改进Hough变换的图像处理方法提高了工件轮廓线检测的精度,并在计算机上对工件参数进行自动评定计算,改进了读数系统的性能.检测结果表明改进后的系统有效地提高了测量的精密度.     

一种基于分道线模型的道路识别方法

介绍了一种基于分道线模型的道路识别方法。该方法首先对图像进行预处理,然后应用分水岭算法划分道路感兴趣区域ROI,最后采用Hough变换拟合分道线,实时地进行道路识别。试验结果显示,在一定条件下,该算法能够满足道路识别的要求。     

概率Hough变换在仪表检测中的应用

介绍了图像处理技术在汽车仪表检测中的应用。利用图像处理技术,如阈值分割、图像膨胀、图像细化等各种图像处理方法进行图像处理。对表盘指针的识别,采用了一种概率Hough变换法。     

SAR图像中机场跑道的自动识别研究

提供了SAR图像中机场跑道自动识别系统，该系统以改进的数学形态学预处理为基础，用改进和扩充的弗里曼链码法初步提取直线，最后用Hough变换精确提取直线，通过对20多幅不同地区、不同波段的图像进行识别，该系统识别准确率达到95％以上，识别时间不超过30s，基本做到了全自动、实时识别．     

Iris Localization Based on Edge Searching Strategies

An iris localization scheme based on edge searching strategies is presented. First, the edge detection operator Laplacian-of-Gaussian （LOG） is used to iris original image to search its inner boundary. Then, a circle detection operator is introduced to locate the outer boundary and its center, which is invariant of translation, rotation and scale. Finally, the method of curve fitting is developed in localization of eyelid. The performance of the proposed method is tested with 756 iris images from 108 different classes in CASIA Iris Database and compared with the conventional Hough transform method. The experimental results show that without loss of localization accuracy, the proposed iris localization algorithm is apparendy faster than Hough transform.     

STUDY ON LIQUID-LIQUID EXTRACTION OF RESVERATROL FROM GIANT KNOTWEED AND ITS PROTECTIVE EFFECT ON MYOCARDIUM INJURY

Ｒｅｓｖｅｒａｔｒｏｌｈａｓｂｅｅｎｆｏｕｎｄｉｎｍａｎｙｐｌａｎｔｓａｎｄｈａｓａｌｏｔｏｆｖａｌｕａｂｌｅｅｆｆｅｃｔｓｕｃｈａｓａｎｔｉ ｔｕｍｏｒ,ａｎｔｉ ｉｎｆｌａｍ ｍａｔｉｏｎ ,ｔｒｅａｔｍｅｎｔｏｆｃａｒｄｉｏ ｖａｓｃｕｌａｒｄｉｓｅａｓｅｓ[1] .Ｔｈｅｍｅｃｈａｎｉｓｍｉｓｒｅｌａｔｅｄｔｏｔｈｅｑｕｅｎｃｈｉｎｇｏｆｌｉｐｉｄｐｅｒｏｘｉｄａ ｔｉｏｎ[2 ,3] .Ｍａｎｙｐａｐｅｒｓｒｅｐｏｒｔｅｄｔｈａｔｒｅｓｖｅｒａｔｒｏｌｗａｓｅｘ ｔｒａｃｔｅｄｆｒｏｍｇｒａｐｅ .Ｈｏｗｅｖ…     

基于混沌遗传算法的基元提取

提出了一种直接在灰度图像上进行基元提取的方法．与传统基元提取方法相比，该方法不需对灰度图像进行边缘检测二值化处理，而是利用图像边缘的梯度信息对混沌遗传算法导向，直接在灰度图像上提取基元．同时，该方法也避免了Hough变换高的时空开销问题．仿真表明该方法方便、有效并具有较高稳定性．     

A method for determining surface free energy of bamboo fiber materials by applying Fowkes theory and using computer aided machine vision based measurement technique

The purpose of this study is to develop a standard methodology for measuring the surface free energy (SFE), and its component parts of bamboo fiber materials. The current methods was reviewed to determine the surface tension of natural fibers and the disadvantages of techniques used were discussed. Although numerous techniques have been employed to characterize surface tension of natural fibers, it seems that the credibility of results obtained may often be dubious. In this paper, critical surface tension estimates were obtained from computer aided machine vision based measurement. Data were then analyzed by the least squares method to estimate the components of SFE. SFE was estimated by least squares analysis and also by Schultz?? method. By using the Fowkes method the polar and disperse fractions of the surface free energy of bamboo fiber materials can be obtained. Strictly speaking, this method is based on a combination of the knowledge of Fowkes theory. SFE is desirable when adhesion is required, and it avoids some of the limitations of existing studies which has been proposed. The calculation steps described in this research are only intended to explain the methods. The results show that the method that only determines SFE as a single parameter may be unable to differentiate adequately between bamboo fiber materials, but it is feasible and very efficient. In order to obtain the maximum performance from the computer aided machine vision based measurement instruments, this measurement should be recommended and kept available for reference.     

COLOR AC-PDP ENHANCEMENT OF LOW LEVELS IMAGE QUALITY WITH DYNAMIC MODIFYING SUB-FIELDS CODE

Based on the ＂Grayscales average distribution＂ method which equally distributes the input gray levels to output gray levels, three improved methods named： ＂Reduce the gray range expressed by the less significant subfields＂, ＂Low levels preset＂ and ＂Modify the exponent of inverse-gamma function＂ are proposed in this paper. Using these methods, the inverse-gamma relation subfields code can be obtained easily which can improve the low level expressions of AC-PDP. And a program, ＂ gray scales distribution validate program＂, which can enhance the expressions of the demanded gray levels range, is also proposed in this paper.     

单向短纤维复合材料中热残余应力研究

基于单向短纤维增强金属基复合材料的单纤维模型,利用轴对称弹塑性有限元分析方法,研究了该类复合材料中热残余应力的大小和分布情况以及各种微结构参数的变化对它的影响,同时还讨论了热残余应力的存在对基体与纤维之间应力传递的影响。研究表明,热应变的不匹配导致纤维受压、基体受拉,界面法向受压。热残余应力的存在降低了基体与纤维间的应力传递,不利于复合材料整体性能的提高。     

Moment-Based Method to Estimate Image Affine Transform

The estimation of affine transform is a crucial problem in the image recognition field. This paper resorted to some invariant properties under translation, rotation and scaling, and proposed a simple method to estimate the affine transform kernel of the two-dimensional gray image. Maps, applying to the original, produce some correlative points that can accurately reflect the affine transform feature of the image. Furthermore, unknown variables existing in the kernel of the transform are calculated. The whole scheme only refers to one-order moment,therefore, it has very good stability.