没有合适的资源?快使用搜索试试~ 我知道了~
人工智人-家居设计-3D智能传感器视觉测量系统关键算法研究.pdf
1.该资源内容由用户上传,如若侵权请联系客服进行举报
2.虚拟产品一经售出概不退款(资源遇到问题,请及时私信上传者)
2.虚拟产品一经售出概不退款(资源遇到问题,请及时私信上传者)
版权申诉
0 下载量 14 浏览量
2022-07-03
10:32:37
上传
评论
收藏 2.1MB PDF 举报
温馨提示
试读
68页
人工智人-家居设计-3D智能传感器视觉测量系统关键算法研究.pdf
资源推荐
资源详情
资源评论
摘要
I
摘 要
3D
智能传感器基于三维结构光视觉测量,它集结构光激光器、传感器和测量
软件于一身,具有运行稳定可靠,维护方便简单,运行和构建成本低廉的特点,
可广泛应用于工业自动化、汽车制造、道路检测等系统中。对
3D
智能传感器测量
系统关键算法的研究,有利于提高系统检测精度,并间接降低检测成本,具有重
要的研究价值。本文主要研究内容分四部分。
1.
摄像机标定算法
研究了摄像机标定的基础内容:摄像机的针孔模型,摄像机的内、外部参数、
镜头的两种畸变模型:除法模型和多项式模型。然后研究相机标定的两类方法,
线性标定方法和非线性标定方法,并深入研究了非线性标定方法中的张正友方法。
利用张的标定方法,进行摄像机标定实验,实验结果与
Opencv
的标定结果一致。
最后分析使用标定图像的数量与标定结果准确度的关系,镜头的光圈和焦距对标
定结果的影响,提出了进行摄像机标定过程中需要注意的事项。
2.
高精度椭圆拟合算法和亚像素光条中心提取算法
通过将椭圆拟合问题转化为非线性最优化问题,获取圆阵列标定板图像高精
度圆心坐标,从而提高摄像机标定结果的精度。对比了常规一阶、二阶像素级微
分边缘算子和亚像素精度边缘检测算法,使用基于
Hessian
矩阵法和提出一种基于
形态学细化方法的结构光光条中心提取算法,并进行光条中心提取实验。实验结
果表明,二者的精度分别为亚像素级和像素级。
3. 3D
智能传感器标定方法
研究了传感器的
3
种测量模型:解析几何模型、面面模型和线面模型;研究
了传感器结构光光平面的几种标定方法,分析它们的优缺点,利用双重交比不变
的标定方法进行系统标定实验,得到了结构光光平面方程参数。最后理论分析影
响结构光测量精度的因素。
4.
关键算法验证软件
介绍系统算法验证软件的摄像机标定模块和系统测量模块;对摄像机标定模
块,介绍了两大组成部分:标定算法模块和相机模块,并介绍了部分核心代码;
最后介绍了摄像机标定模块的执行流程。
据我们所知,采用基于形态学细化原理进行结构光光条中心提取方法,国内还
未见报道。
关键词:3D 传感器,摄像机标定,结构光光平面标定,光条中心提取,椭圆拟合
ABSTRACT
II
ABSTRACT
3D smart sensor, which based on structured light vision measurement, combines
structured light laser, sensor and measurement software in one. Since it has the
characteristics of reliable running, convenient maintenance, simple operation and low
cost of building, can be widely used in industrial automation, automobile manufacturing
and road detection systems. In this thesis, the key algorithms of 3D smart sensor
measuring system were studied. The study of key algorithms of 3D smart sensor, which
is beneficial to improve the system precision and indirectly reduce the test cost, have
important research value. The main research contents the next four parts.
1. The camera calibration algorithm
The basic content of camera calibration is studied: the pinhole camera model, the
internal and external parameters of camera, two kinds of lens distortion model: division
model and polynomial model.
Then the two types of camera calibration methods were
researched: linear and nonlinear calibration methods, and the further study of Zhang
Zhengyou calibration method was made. By using the Zhang Zhengyou calibration
method, we got that the experimental results consistent with the calibration results of
Opencv. Finally, the relationship of number of images and the calibration accuracy, the
influence of lens’ aperture and focal length on the calibration result were analyzed, the
matters which need to pay attention to in the process of camera calibration was puts
forward.
2. The high precision ellipse fitting method and the sub-pixel structure light center
extraction algorithm
In order to get the circular array center of calibration plate image with high
precision, transforming ellipse fitting into a nonlinear optimization problem, so as to
improve the precision of camera calibration results. Then conventional pixel level first
and second order differential edge operator, sub-pixel edge detection algorithms were
compared with each other. Using a method based on Hessian matrix and putting forward
a kind of algorithm based on morphology thinning to extract center of structure light.
Experimental results show that the accuracy of the two algorithms was sub-pixel level
and pixel-level respectively.
3. 3D smart sensor calibration methods
ABSTRACT
III
At first, measuring principle and three kinds of measurement models of 3D smart
sensor was studied. Then several structure light plane’s calibration methods and their
advantages and disadvantages were analyzed. Through the experiment calibration of 3D
smart sensor, laser light plane equation parameters were obtained.
Finally the influence
factors of affecting the structure light measurement were analyzed in theory.
4. Verification software of key algorithms
The camera calibration module and the measurement calibration module of system
software were introduced at first. Then In view of the camera calibration module, this
paper introduced two major parts: camera calibration algorithm modules and camera
modules, and some core codes were introduced. Finally the camera calibration module’s
implementation process was introduced.
To the best of our knowledge, the method which based on morphology thinning
principle to extract center of structure light strip, has not been reported in China.
Keywords:
3D smart sensor, camera calibration, the structure light plane calibration,
center extraction of structure light, ellipse fitting
目录
IV
目 录
第一章 绪 论 .................................................................................................................. 1
1.1 引言 .................................................................................................................... 1
1.2 机器视觉 ............................................................................................................ 1
1.2.1 机器视觉概述 .......................................................................................... 1
1.2.2 机器视觉的应用领域及面临的问题 ...................................................... 2
1.2.3 视觉测量 .................................................................................................. 3
1.3 三维测量技术概述 ............................................................................................ 3
1.3.1 接触式测量技术 ...................................................................................... 4
1.3.2 非接触式测量技术 .................................................................................. 4
1.3.2.1 主动式测量技术 ............................................................................ 4
1.3.2.2 被动式测量技术 ............................................................................ 5
1.4 3D 智能传感器视觉测量系统关键算法 ........................................................... 6
1.5 本文结构 ............................................................................................................ 7
第二章 摄像机标定算法研究 ........................................................................................ 9
2.1 相机模型及其参数 ............................................................................................ 9
2.1.1 面阵相机与针孔相机模型 ...................................................................... 9
2.1.1.1 相机外参数 .................................................................................. 10
2.1.1.2 镜头畸变 ...................................................................................... 10
2.1.1.3 相机内参数 ................................................................................... 11
2.2 摄像机标定 ...................................................................................................... 12
2.2.1 线性标定法 ............................................................................................. 12
2.2.2 非线性标定法 ......................................................................................... 13
2.2.3 两步标定法 ............................................................................................. 13
2.3 摄像机标定实验及精度评价 .......................................................................... 18
2.3.1 实验准备 ................................................................................................. 18
2.3.2 实验过程 ................................................................................................ 18
2.3.3 标定结果 ................................................................................................ 19
2.3.4 标定参数的准确度 ................................................................................ 20
2.4 本章小结 .......................................................................................................... 23
第三章 高精度图像特征提取及光条中心提取算法研究 .......................................... 24
目录
V
3.1 像素级精度边缘检测 ...................................................................................... 24
3.1.1 一阶微分边缘检测算子 ........................................................................ 24
3.1.1.1 Roberts 边缘算子 .......................................................................... 24
3.1.1.2 Sobel 边缘算子 ............................................................................. 25
3.1.1.3 Canny 边缘算子 ............................................................................ 25
3.1.2 二阶微分边缘检测算子 ........................................................................ 26
3.2 亚像素级精度边缘检测 .................................................................................. 27
3.3 高精度椭圆拟合 .............................................................................................. 28
3.3.1 重心法 .................................................................................................... 28
3.3.1.1 带阈值的重心法 .......................................................................... 28
3.3.1.2 平方加权重心法 .......................................................................... 29
3.3.2 高精度拟合法 ........................................................................................ 29
3.3.3 圆阵列中心提取实验 ............................................................................ 31
3.4 光条中心提取算法研究 .................................................................................. 32
3.4.1 基于形态学细化的光条中心提取方法 ................................................ 32
3.4.2 亚像素光条中心提取方法 .................................................................... 34
3.5 光条中心定位结果及精度 .............................................................................. 36
3.6 本章小结 .......................................................................................................... 37
第四章 智能传感器标定方法研究 .............................................................................. 38
4.1 测量原理 .......................................................................................................... 38
4.1.2 结构光激光器 ........................................................................................ 39
4.2 结构光三维视觉测量模型 .............................................................................. 39
4.2.1 解析几何模型 ........................................................................................ 39
4.2.2 透视投影测量模型 ................................................................................ 42
4.2.2.1 线结构光视觉面面测量模型 ...................................................... 42
4.2.2.2 线结构光视觉线面测量模型 ...................................................... 44
4.3 测量系统标定方法 .......................................................................................... 45
4.3.1 锯齿靶法 ................................................................................................ 45
4.3.2 微分法 .................................................................................................... 46
4.3.3 已知直线(平面)和图像点匹配法 ......................................................... 47
4.3.4 基于双重交比不变的标定方法 ............................................................ 47
4.3.4.1 双重交比不变原理 ...................................................................... 48
4.3.4.2 标定靶标的设计 .......................................................................... 49
剩余67页未读,继续阅读
资源评论
programmh
- 粉丝: 4
- 资源: 2163
上传资源 快速赚钱
- 我的内容管理 展开
- 我的资源 快来上传第一个资源
- 我的收益 登录查看自己的收益
- 我的积分 登录查看自己的积分
- 我的C币 登录后查看C币余额
- 我的收藏
- 我的下载
- 下载帮助
最新资源
- Unity3D游戏源码PLUS教程(2018)3D传送门
- 管理系统源码是一套用于支持和优化组织内部运营的软件程序资源,涵盖了从项目规划、执行到监控的全方位管理功能 以下是关于管理系统源码
- Unity3D游戏源码PLUS教程火炬之光demo第四代
- Python爬虫是一种使用Python编程语言来自动化获取网页数据的技术 这项技术主要涉及到向目标服务器发送请求,获取HTML页
- Unity3D游戏源码PLUS教程大型ARPG手游(降临)客户端服务端全套源码
- 蓝桥杯,作为全国软件和信息技术专业人才大赛的知名品牌,旨在促进软件和信息技术领域专业人才培养、推动产学研用有机结合、提高大学生的
- Unity3D游戏源码PLUS教程CS源码(支持局域网+广域网)
- 智能车,作为现代汽车科技的重要发展方向,集合了环境感知、规划决策、多等级辅助驾驶等多种功能于一体,是计算机、现代传感、信息融合、
- Unity3D游戏源码PLUS教程(众神之光)卡牌u3d全套完整源码+资源
- Unity3D游戏源码PLUS教程(口袋三国)客户端+服务器+设计文档
资源上传下载、课程学习等过程中有任何疑问或建议,欢迎提出宝贵意见哦~我们会及时处理!
点击此处反馈
安全验证
文档复制为VIP权益,开通VIP直接复制
信息提交成功