字符识别

The system frame and workflow of our LPR System are introduced. Some problems and two of our proposed character recognition methods, i.e. method based on PCA-LSM for limited Chinese character recognition and method based on structural feature analysis for alphabetic and digital character recognition, are addressed in full details.

文中介绍了该系统的结构及工作流程，以及两种字符的识别方法：基于PCA-LSM的有限中文字符识别方法和基于结构特征分析的字母及数字字符识别方法。

This paper focus on the study of the probability based extraction method which includingword, text line and text block extraction in our Optical Character Recognition System. Inpre-process of OCR, several methods of the two problems:image binarisation are alsodiscussed, thus corresponding methods are finally selected and determined.

本文主要对字符识别系统中基于概率的提取方法进行了研究，其中包括对文本行的提取、块的提取和词的提取，并对字符识别预处理中的两个问题：图象二值化和倾斜字体校正的几个方法的优劣进行了讨论和试验模拟，并最终为我们的字符识别系统确定了相应的方法。

3A fast and efficient algorithm is presented, which first thins letter images, then uses improved Hausdorff distance to measure the degree of similarity between templates and letter images.

在车牌字符识别中，为了提高字符识别的稳定性，本文提出了基于改进的Hausdorff距离字符识别算法。

In the layout analysis, we chose a bottom-up algorithm based on nearest neighbor connect-strength and line confidence to segment the image area, table area and text area; In the printed Chinese character recognition module, we calculate the degree of incorporative difference to match the character, also define a refusal class which could orientate the mathematical expression automatically; Put the expression into the formula processing module, then we chose the character segmentation method based on connectivity and the template match method to recognize the character in the mathematical expression and at last we used the structure analysis based on the character to transform the two dimension formula into one dimension Word EQ expression.

在版面分析中，采用基于最近邻连接强度和行列可信度的自底向上的版面分析算法，分割出图像区域、表格区域和文本区域；在汉字识别模块中，采用回溯切分方法切分出字符段，计算合并差异度与特征字典比较，通过引入汉字的拒识类，从而实现了公式的定位；将定位后的数学公式送入公式识别器，在公式识别器中采用基于连通域搜索的字符分割方法和模板匹配方法对字符识别，对于识别出的字符，再采用基于特征字符的结构分析方法，从而将二维的数学公式转化为一维的Word EO域语句。

The authors use the projection method to solve the segmenting problem of the characters in the case of character adhered or intersected with the lines, which made the recognition accuracy of the omnirange string increased.

运用波形投影方法解决了粘连字符及字线相交情况下的字间切割问题，使工程图多向字符识别精度显著提高，该算法对局部退化状态下的字符识别具有良好的抗干扰性。

Firstly, the system is divided into four sub systems: Chinese characters recognition system, character recognition system, recognition system including characters and digits, and digital recognition system; Secondly, the statistic information about left scan line free length of every character is used as a group feature of car plate, which will accomplish the recognition together with statistic template by weighted sum.

首先将车牌字符识别分成四个子识别系统——汉字识别系统、字母识别系统、字母和数字混和识别系统，以及数字识别系统；然后选择各字符的左扫描线空程长度统计信息作为车牌字符的一组特征，与统计模板进行加权组合，来共同完成车牌字符的识别工作。

General characters recognition algorithms can't offer a good way to recognize characters of deflective license plate, and before they recognize deflective characters, a step named deflection rectifying must be progressed, but the effect of this step is a big influence on the result of characters recognition.

一般的字符识别算法都不能很好的解决倾斜车牌字符的识别，在识别倾斜字符时都要先经过倾斜校正这一步，而倾斜校正的效果往往对最后的识别结果有很大的影响。

On the license location,the projection was used to locate the license plate; On the characters segmentation, the liminal rule was used to divide the characters; In order to solve the problem of the digital characters recognition in the plate, BP nerve network was used to recognize the digital characters.

本文对车牌识别系统中的车牌定位、字符分割和字符识别进行了初步研究。对车牌定位，本文采用投影法对车牌进行定位；在字符分割方面，本文使用阈值规则进行字符分割；针对车牌图像中数字字符识别的问题，本文采用了基于BP神经网络的识别方法。

This article comminutes the symbols of the formula and records their position by the outline rectangles of symbols, and then normalized these symbol. Then designed Artificial Neural Network that recognizes these symbol.

本文通过计算公式中字符的最小外接矩形，将字符依次分割开并记录其位置信息，然后进行归一化处理，利用本文设计的适合本文中字符识别的识别模块，即二层感知器神经网络识别模块，将归一化后的字符识别出来。

This system completed character segmentation part using the similarity between a character string and license plate algorithm, and made corresponding improvement, and completed character recognition with wavelet transformation and support vector machine.

LPR包括字符切分和字符识别两部分，本系统用字符串车牌相似度的方法完成字符切分；用小波分解与支持向量机来识别车牌字符。

Breath, muscle contraction of the buttocks; arch body, as far as possible to hold his head, right leg straight towards the ceiling (peg-leg knee in order to avoid muscle tension).

呼气，收缩臀部肌肉；拱起身体，尽量抬起头来，右腿伸直朝向天花板（膝微屈，以避免肌肉紧张）。

The cost of moving grain food products was unchanged from May, but year over year are up 8%.

粮食产品的运输费用与5月份相比没有变化，但却比去年同期高8％。