For the complex scenes ’ wiping off , firstly , the algorithm of feature fuzzy matching is applied to wipe off most of the still and luminous insensitive scenes , then , a kind of area connecting technology is used to wipe off the rest scenes . at last , the vehicle body area which has been mismatched will be repaired 对多车牌图像的复杂背景去除,本文首先利用特征模糊匹配法将大部分静止且感应光线变化不敏感的背景去除,然后利用图像区域连通技术将残留未去除的背景进一步消除,并对误匹配的车体区进行修复。
There are three steps in the combined . firstly , the template matching method is used to classify roughly in the first step ; secondly , the fuzzy pattern recognition method which are connected paralled is used to classify finely ; at last , the third classifier , which input is the output of the second classifiers , raise one ’ s hand as its decision rules 第一级分类器采用模板匹配法进行粗分类;第二级分类器由并行的模糊模式识别分类器组成,进行细分类;第三级分类器是综合分类器,它将第二级的输出作为输入,根据举手表决规则得到最后的输出结果。
To the complete character use the image which representation by matrix , making the subtraction between the matrix of capture image and the matrix of template , the result matrix which absolute sum of its all elements is the smallest is the recognition result . to the incomplete character , using two lines of 0 ~ 9 as template , then carry on the recognition using the template matching . to the scale division recognition , using the central position of it , distinguishes the long scale division expression is merely 0 perhaps 5 对完整字符,采用将图像用矩阵表示,利用获取图像的矩阵与识别模板的矩阵做减法,结果矩阵的所有元素绝对值和最小的即为识别结果;对不完整字符,针对电度表读数的具体特点,采用将两行0 ~ 9的字符图像排列起来作为模板,利用模板匹配法进行识别;对于刻度的识别,利用刻度的中心坐标位置,仅仅识别长刻度表示的是0或是5即可。
Then based on the idea of predictive motion vector , using of spatial correlation of adjacent block and global minimum points probability distribution characteristic , predictive diamond searching ( pds ) and its advanced mode : adaptive pds ( apds ) are introduced . finally the algorithm of pds and apds and its simulation results comparing with conventional me algorithm are given 然后基于预测性运动矢量的概念,利用相邻块运动矢量的相关性和全局极值点概率分布特性,提出了预测性菱形搜索算法和它的改进算法:自适应预测性菱形搜索法,设计出具体算法,并给出了与传统快速块匹配法比较的计算机仿真结果。
Among them the gray level co - occurrence matrix ( glcm ) and gray gradient co - occurrence matrix ( ggcm ) methods , which attributed to the statistic textural analysis scheme were then chosen to extract the textural features of five kind areas on satellite images . in the second part the principle of classification and bp neural network were introduced . combined with textural features , the improved bp neural network successfully performed on the classification of the satellite images 论文的第一部分介绍了进行纹理特征研究的一些典型的方法,利用其中的基于统计的纹理分析法中的灰度共生矩阵以及灰度一梯度共生矩阵法,分析了卫星云图上五类区域的纹理特性;第二部分主要介绍了遥感图像分类原理以及神经网络中的bp算法,在对算法原理进行深入理解的基础上,把纹理特征与神经网络进行组合,实现对卫星云图进行分类分析;第三部分内容是在前面图像分类结果的基础上,对序列图像用相关匹配法进行运动分析,反演云迹风风场。
According as the base principle of digital video motion parameter estimating , placed emphasis on discussing the block - matching method in motion estimation , gave a intact n - search method , simulated in computer according to the s - vga standard of digital video , and got a accurate motion vectors 依据数字视频运动参数估计的基本原理,重点讨论了运动参数估计中最常用的块匹配法,给出了一个完整的n步搜索法方案。以数字视频演播标准s - vga为标准,采用计算机仿真得到了正确的运动矢量。
Broadband source location method is investigated by using signal - phase - matching principle , and simulations and experiments are carried out to show the validity of the method ; two high resolution algorithms of broadband signals are proposed , which are the dft interpolation algorithm and the frequency invariant beam space algorithm . the proposed algorithms have no requirement for prior information of source locations , so the results of doa estimation can be unbiased ; a method of temporal - spatial averaging , which can improve the performance of doa estimation for uniform linear arrays , is given . this method can be applied in the most narrowband or broadband eigen - structure based high resolution algorithms 提出了利用信号相位匹配法的宽带信号方位估计方法并从仿真和实验两方面对其进行了验证;提出了基于dft插值的和基于频率不变响应波束域的两种宽带信号高分辨方位估计算法,这两种方法在进行方位估计时均不需要进行方位预估,从而避免了因角度预估可能带来的估计偏差;最后提出了一种提高均匀线列阵方位估计性能的时空平均法,该方法可适用于多数窄带或宽带的特征结构类高分辨方位估计算法中。
Detaching the wear region and background , caculating the threshold value is 151 , 147 , 118 taking overlap method and maximum dispersion method of high - load small particle abrasive wears low - load big particle abrasive wear and adhesive wear , then processing edge detection and skeleton map with classical sobel and gauss - laplace and kirsch arithmetic operators of template matching method , making the shape skeleton produced by different wear mechanism embodied cleared . and analysing the trait of shape skeleton 然后应用边缘检测的经典边缘检测算子sobel算子、 gauss - laplace算子和模板匹配法中的kirsch方法对图像进行边缘检测以及轮廓提取,使这三种磨损形式的不同磨损机制所产生的形状轮廓得到清晰地体现,并结合磨损机制分析形状轮廓的特点。
A method is proposed for the parameter identification of the viscoelastic internal dampers of the multirotor systems in this paper . the multi - rotor system is regarded as a composite structure . the impedance matching method in the substructure method for structural dynamic analysis is used for the dynamic calculation of the structure . the augmented lagrange method is applied to the iteration calculations of the characteristic determinant of eigenvalve equation of the system . and then the results of parameter identification can be obtained . upon the dynamic characteristics of the system , the method can reveal the influence of the gyroscopic moments of force precisely . in process of the analysis , the various properties of multi - rotor systems are made the best use of , so the measure of the modes parameters is avoided , the calculation is simplified . in addition , an efficient method is proposed to establish a dynamic model of multirotor systems with viscoelastic internal dampers . the calculating results are consistent with the experimental results satisfactorily 提出了多转子系统的粘弹性中间弹支参数辨识的一种方法.该方法把多转子系统看作复合结构,用动态子结构分析的阻抗匹配法计算动力学特性,用增广拉格朗日乘子法迭代计算系统特征方程的系数行列式,获得系统物理参数的辨识结果.该方法可以准确计入陀螺力矩的影响,避免测量振型参数.在分析中,利用多转子系统的特性可简化分析过程.文中还介绍了一种建立粘弹性弹支多转子系统动力学模型的方法.计算结果与试验结果基本一致
Firstly , the background and the significance of the project are introduced and the current research of related fields is reviewed . secondly , the sales work - flow of mold & die is analyzed from the point of view of customers . considering the characteristics of different stages of the whole process , different searching or evaluating methods are adopted , i . e . , maximum requirement matching , fuzzy - ahp and hybrid conjoint analysis , thus , the decision - making support is provided for the customer 然后以模具行业为背景,从客户的角度对客户、模具企业共同参与的互动的模具营销过程进行分析,并针对从招标到签约整个过程中各个阶段的不同特点,采用不同的搜索或评估方式,即分别在招标阶段、询价阶段和签约阶段采用极大需求匹配法、模糊层次分析法、建立自明确模型的方法,为客户的最终的订单决策提供支持。