摘 要

针对土工试验室土壤环刀样品天平称重需要人工记录环刀编号、耗费时间与人力的问题，研究了图像识别在土工实验自动化称重中针对环刀钢印编号的应用。

论文主要研究分析了环刀钢印编号识别研究的背景以及现状，提出了环刀钢印编号识别系统的总体设计方案，设计了一种基于penCV（计算机视觉库）和Tesseract识别训练引擎相结合的环刀编码识别系统，该系统硬件采用LifeCam Studio高清网络摄像机作为图像数据采集平台，结合软件平台Windows操作系统实现了图像采集、图像处理以及字库训练等功能。在环刀钢印编号的自动定位上利用基于HSV颜色空间的自动定位算法，有效的消除了环刀外壁由于长期与土壤、泥沙接触而形成的灰色磨损印记所带来的影响，完成了钢印编号的精准定位；利用开源OCR识别引擎Tesseract对预处理得到的图像进行识别和训练。Windows平台下的识别系统的搭建包括利用CMake和Visual Studio2015对OpenCV源代码和Tesseract源代码的编译生成相应的静态库与动态链接库，利用Visual Studio2015在设计的MFC应用程序中调用相应的动态链接库，实现对环刀钢印编号的识别与训练，经过运行测试，环刀钢印编号识别系统运行稳定，能迅速的识别相应的数字编号，通过训练模式，可以加快识别的速率，提高工作效率。

研究结果表明该系统采取的定位、识别、训练算法能够满足环刀钢印编号识别精度、准确度和响应时间的要求，实现了钢印编号的自动识别与处理，极大地方便了土工试验人员开展相应的土工试验。

关键字：环刀钢印编号；OpenCV；Tesseract

Abstract

Presently, in the field of geotechnical experiment, due to the fact that many test operation process are more complicated, test personnel need to sampling the experimental data by themselves. Thus, there are many shortcomings in the process, such as inefficient, low automation and low intelligence, which requires a lot of automated instrument design, in order to improve the production efficiency and shorten the construction period of the project of construction. This paper studies the application of image recognition in the automatic weighing of geomembrane of the geometric experiment, that is to recognize the steel seal number on the cutting ring.

This paper mainly analyzes the background and current situation of the research on the identification of the steel stamps on the cutting ring, and puts forward the overall design scheme of the steel seal number of the cutting ring recognition system, and design a smart steel seal number Recognition System based on the OpenCV cross-platform computer vision library and Tesseract recognition amp; training engine. The hardware of this system use the Microsoft’s LifeCam Studio HD video camera as an image data acquisition platform, combined with the software platform which based on the Windows operating system to achieve the image acquisition, image processing, image recognition and font train and other functions. The automatic positioning algorithm based on the HSV color space is used to automatically eliminate the influence of the gray mark formed by long-term contact with the soil and sediment. In the the way of character segmentation and recognition, using the open source recognition engine—Tesseract for processing recognition and training. In the Windows platform, We use the software of CMake and Visual Studio 2015 to compile the source code of the OpenCV and Tesseract to the static library and dynamic library and using the VS 2015 complier to create the MFC application by using the corresponding dynamic link library. Through the test, the steel seal number of the cutting ring recognition system is stable, and it can quickly identify the corresponding number, through the training mode, it can speed up the recognition process, improve the work efficiency.

The results show that the positioning, recognition and training algorithms can meet the requirements of the accuracy and response time of the number of cutting ring, and realize the automatic identification and processing of the stamp number, which greatly facilitates the geotechnical test personnel to make the experiment of the geotechnical test

Key Words: the steel seal number of cutting ring; OpenCV; Tesseract.

目录

第1章 绪论 1

1.1课题背景及意义 1

1.2 国内外研究现状 3

1.3 本论文主要研究内容 4

第2章 环刀编码识别系统总体方案设计 6

2.1 环刀编码识别系统功能分析 6

2.2 识别引擎选择 7

2.3 识别系统方案论证 8

2.4 识别系统硬件组成 9

2.5 系统软件框架及主要流程 9

第3章 环刀编码识别系统硬件设计 11

3.1 图像输入设备的选型 11

3.2 系统光源系统的选择 14

3.3 硬件设备成本分析 16

第4章 环刀编码识别系统算法设计 18

4.1 字符区域自动定位算法分析 18

4.1.1 RGB颜色空间模型 20

4.1.2 HSV颜色空间模型 20

4.1.3 RGB颜色空间转到HSV颜色空间算法 23

4.2 灰度图二值化算法分析 35

4.3 滤波算法分析 40

4.3.1 中值滤波算法简介 41

4.3.2 中值滤波算法过程 41

4.4 识别训练方法分析 42

第5章 环刀编码识别系统软件设计 44

5.1 OpenCV简介 44

5.2 Tesseract简介 44

5.3 OpenCV源代码的编译 48

5.4 Tesseract源代码的编译 50

5.5 MFC工程的建立 51

5.6 本章小结 55

第6章 系统功能测试 57

6.1 系统测试环境与方法 57

6.2 系统功能测试 57

6.3 系统功能测试结果分析 62

第7章 总结与展望 64

7.1 全文总结 64

7.2 展望 64

参考文献 66

附录A MFC设计主要代码 67

附A1 MainFrm.cpp 67

附A2 CvvImage.cpp 71

附A3 ChildFram.cpp 80

附A4 BoxProcWnd.cpp 82

附A5 SKTiffHelper.h 84

附A6 TestRecoDoc.cpp 88

附A7 TestRecoView.cpp 91

附A8 MFCApplication1.cpp 99

附A9 MFCApplication1Doc.cpp 106

附A10 MFCApplication1View.cpp 110

致谢 115

第1章 绪论

1.1课题背景及意义

土工试验就是对土壤等的物理化学、力学性质以及成分进行测试分析的过程，按照测试的地点不同，土工试验可以分为在土样采集现场进行测试和在试验室内进行测试这两种方式^[11]。土工试验为建筑工程项目前期的勘查测试提供相关土壤试样的精确测试数据，从而根据相应的试验结果，判定该区域的地质条件是否满足建筑施工的条件，是否符合相应的建筑标准，更进一步的为建筑项目设计解决方案。在实验室内进行测试时，包括对土壤样本进行含水量、容重、比重等试验。在进行土壤含水量测试时，需要采集土样，而在采集土样时，必不可缺的取样工具就是环刀，它的材质一般为不锈钢，环刀外形如图1.1所示。