搜索详情-毕业论文网

注册

  • 获取手机验证码 60
  • 注册

找回密码

  • 获取手机验证码60
  • 找回

微小图像的自动识别算法与软件的实现毕业论文

 2021-02-28 21:37:00  

摘 要

视觉通道是人类获取外界事物的主要通道,图像则是承载大量信息的最为直观的表现形式。随着计算机技术在各个领域的迅猛发展,基于计算机的图像识别技术也得到了日益重视并且不断地应用到更广泛的实际工程之中。由于实际工程中对微小图像识别的精度和速度要求越来越高,传统的识别方法就难以符合要求了。bar条上的激光器图像只有在显微镜等高倍率光学设备下才可观察到,人工无法直接完成相应的生产测试,所以激光器编码字符识别的研究对生产测试有着重要的意义。现阶段,国内外对字符识别技术的研究已经趋于成熟,但是对bar条上激光器编码字符的研究还比较少。

本文将以bar条上激光器阵列上的字符作为微小图像的研究对象,对激光器图像的主要特征和编码识别进行分析,设计算法能够自动识别激光器阵列上的微小图像,在vs2010开发环境和opencv2.4.9搭建系统测试平台对该系统性能进行测试。

针对目前的研究情况,对一些关键技术进行分析和研究,主要解决以下几个方面的问题:

(1)如何准确快速地定位分割半导体激光器阵列;

(2)如何精确地切分出每个编码字符图像;

(3)怎样设计高识别率的识别器;

(4)规划系统设计和需要实现的功能。

因此,为解决上述问题,本文将从以下几个方面进行系统的设计:

(1)图像预处理:先对采集到的图片进行灰度化处理,然后通过空域平滑滤波的方式进行图像去燥,最后通过直方图均衡法进行图像增强;

(2)激光器的定位分割:先基于Sobel算子进行边缘提取,再进行二值化处理,然后利用边缘点密度方式形成连通区域,最后根据轮廓跟踪算法提取轮廓,完成图像的分割;

(3)字符分割:在水平、垂直方向对边缘二值化图像进行投影分析,结合激光器突出特征,利用最大方差的单字符切分方法完成对字符的高精度提取;

(4)字符识别:先进行字符特征提取,然后通过特征模板匹配的方法进行字符识别。

关键词:图像预处理;图像定位分割;字符切分;字符识别

Abstract

Visual channel is the main channel for human beings to get outside things, and the image is the most intuitive form of carrying a lot of information. With the rapid development of computer technology in various fields, computer-based image recognition technology has received increasing attention and has been applied to a wider range of practical projects. As the accuracy and speed of the small image recognition in the actual project is getting higher and higher, the traditional method of identification is difficult to meet the requirements. The laser image on the bar is only observable under the microscope with high magnification optical equipment, and the corresponding production test can not be done directly. Therefore, the research of laser coding character recognition is of great significance to the production test. At present, domestic and foreign research on character recognition technology has become more mature, but the study of laser dance code characters is still relatively a little.

In this paper, we will use the characters on the laser array of the laborers as the research object of small images. The main features and coding recognition of the laser image are analyzed. The algorithm can automatically identify the tiny images on the laser array in the vs2010 development environment and opencv2. 4.9 build system test platform to test the performance of the system.

In view of the current research situation, some key technology analysis and research, mainly to solve the following aspects:

(1) how to accurately and quickly locate the segmented semiconductor laser array;

(2) how to accurately cut out each encoded character image;

(3) how to design a high recognition rate of the recognizer;

(4) planning system design and the need to achieve the function.

Therefore, in order to solve the above problems, this paper will be from the following aspects of the system design:

(1) image preprocessing: first of the collected image gray processing, and then through the airspace smoothing filter to dry the image, and finally through the histogram equalization method for image enhancement;

(2) the localization of the laser segmentation: first based on the Sobel operator edge extraction, and then binarization processing, and then use the edge point density to form a connected area, and finally according to the contour tracking algorithm to extract the contour to complete the image segmentation;

(3) character segmentation: in the horizontal and vertical direction of the edge of the binarized image projection analysis, combined with the laser highlight features, the use of the largest variance of the single-character segmentation method to complete the high precision extraction of characters;

(4) character recognition: first character feature extraction, and then through the feature template matching method for character recognition.

Key Words: Image preprocessing; image localization segmentation; character segmentation; character recognition

目 录

摘 要 III

Abstract IV

第1章 绪论 1

1.1 研究背景与意义 1

1.2 国内外研究现状 1

1.3 研究内容 2

第2章 算法分析与设计 4

2.1 图像预处理 4

2.1.1 灰度化 4

2.1.2 图像去噪 4

2.1.3 图像增强 5

2.2 激光器图像分割 6

2.2.1 基于边缘检测的分割 7

2.2.2形成连通组件 8

2.2.3 外轮廓提取与分割 9

2.3 编码文字定位及字符切分 10

2.3.1 倾斜校正 10

2.3.2 编码文字定位 11

2.3.3 边缘图像二值化 11

2.3.4 水平投影 12

2.3.5 垂直投影 13

2.3.6 编码区域切分 14

2.4 字符分割 15

2.5 字符识别 15

2.5.1 模板匹配概述 15

2.5.2 特征提取 16

2.5.3 模板库的建立 17

2.5.4 字符分类识别算法 18

第3章 软件实现 19

3.1 界面设计 19

3.2 C 代码编写 20

第4章 系统功能测试与结果分析 21

4.1 系统测试 21

4.2 性能测试 22

4.3 结果分析 23

4.4 存在的不足 23

第5章 总结与展望 25

5.1 工作总结 25

5.2 工作展望 25

参考文献 26

致 谢 27

第1章 绪论

1.1 研究背景与意义

图像识别技术是计算机对不同形式的目标和对象进行处理、分析和识别的过程。识别过程包括:预处理、定位分割、特征提取、判别匹配。简单来说,图像识别技术就是一种计算机能够像人一样读懂图片内容的技术。通过图像识别技术,我们不仅能够通过图像搜索来更快地获取信息,而且还可以与外部世界产生一种新的交互方式,甚至使外部世界更加智能化[1]。随着计算机技术的迅猛发展,图像识别技术也得到了广泛的关注和重视,百度李彦宏在2011年就有提到“全新的读图时代已经到来”,越来越多的公司都开始涉及这一领域,这标志着智能化时代已经悄然而至。

本课题以bar条上的激光器阵列上的编码字符作为微小图像的研究对象,通过图像识别技术定位到激光器阵列上的编码字符所在位置,然后对其进行分割,通过可行的识别技术对编码字符进行识别。

激光器阵列的特点如下所示:

您需要先支付 80元 才能查看全部内容!立即支付

课题毕业论文、开题报告、任务书、外文翻译、程序设计、图纸设计等资料可联系客服协助查找。