| 强辐射环境下基于图像与点云融合的视觉系统研究 | |
| 刘成安 | |
| Thesis Advisor | 李和平 ; 李众立 |
| 2013 | |
| Degree Grantor | 中国科学院研究生院 |
| Place of Conferral | 北京 |
| Degree Name | 博士 |
| Degree Discipline | 环境科学 |
| Keyword | 强辐射环境 抗辐射加固 视觉系统 算法优化 |
| Abstract | 强辐射环境广泛存在于核电站、民用辐照厂、放射性物质存放和处置场地以及外太空等场所。鉴于强辐射对人体健康的严重危害，这类场所中相关设施的日常运行、维护、事故应急处理等作业主要是由机器人或遥控操作完成，准确的地形和位置等现场信息是机器人或遥控操作相关作业的基础。这些信息的获取主要依靠视觉系统完成，但是由于强辐射也会造成对电子设备的损伤，对视觉系统的建立提出了更高的要求。随着核工业、航天航空等行业的发展，强辐射环境下视觉系统的相关理论和技术研究逐渐成为现代科学技术研究领域亟待解决的热点问题之一。 本文围绕核能与核辐照等领域生产需要，针对强辐射环境下视觉系统的辐射防护、数据处理、目标位置测量和信息融合等问题，进行了基于图像和点云融合的视觉系统相关理论和关键技术研究，取得了以下成果： 1、构建了强辐射环境下视觉系统的硬件和支撑软件实验平台。在分析现有视觉系统特点和强辐射环境对系统要求的基础上，提出利用摄像机和3D激光测距仪互补的特点，以它们的信息融合作为系统的视觉模式，并根据强辐射环境下视觉系统任务需要，对硬件组成和配置参数进行了详细的分析，给出了系统的配置方案，构建了一个硬件系统和支撑软件平台。 2、研究了视觉系统的辐射损伤机理并开发出有效的抗辐射屏蔽装置。基于对电子系统的抗辐射加固方法研究现状及发展趋势的分析，对强辐射环境对电子系统的损伤机理，以及激光测距仪、摄像机等设备的耐辐照性能进行了研究：在1.25兆电子伏特能量级、21万居里源活度的60Co辐照环境下，通过对摄像机、控制电路的抗辐射性能实验研究，确认视觉系统中摄像机、单片机、晶闸管等器件的损坏主要是由辐照累计量决定的——摄像头损坏累积辐照剂量约340 Gy、单片机损坏时的辐照累积剂量约为580 Gy，晶闸管损坏时的累积剂量为1330 Gy。进而提出采用铅金属和铅玻璃结合构建抗辐射屏蔽体的防护方案，并设计了一个视觉系统的屏蔽体，在该屏蔽体防护下摄像机和激光测距仪及其控制电路可以正常工作，并且工作时间可以延长2～4倍。 3、提出强辐射环境下的图像和点云数据预处理算法。基于对强辐射环境下系统所采集图像和点云数据噪声的组成和特点、各种滤波算法及其特性的分析，提出以中值滤波和小波变换相结合的滤波算法来降低图像噪声；对于点云数据，提出通过控制滑动窗口进出数据方式、减少比较次数来提高滤波速度的改进型中值滤波算法。最终，通过实验验证了这两种改进算法的有效性。 4、提出了图像与点云角点的提取算法。针对图像角点的提取，提出了一种基于Harris角点提取的改进算法。在MATLAB 7.0环境下对Harris算法与改进算法的稳定性、抗噪声能力进行的仿真实验表明，改进算法的稳定性和抗噪声能力明显优于Harris算法；针对点云角点的提取，分析研究了点云曲率与点云角点的关系，提出了三维点云曲率的计算方法和基于曲率的点云角点提取算法。 5、确定了实现强辐射环境下定位图像数据与点云数据的融合新方法。通过实验研究确定了激光测距仪在强辐射环境下的系统误差标定方法；以黑白棋盘格为目标靶、利用MATLAB标定工具箱和OpenCV标定工具箱对摄像机内部参数和畸变参数进行了标定；通过对像素坐标与点云坐标的几何约束关系进行分析，建立了点云坐标与摄像机内部参数和像素坐标的数学模型；提出了采用一块镂空的棋盘格作靶标和基于角点为特征点的摄像机和激光测距仪联合标定的新方法。采用以上方法，通过联合标定实验建立起系统的旋转和平移矩阵，实现了图像数据与点云数据的融合。 6、提出了基于RANSAC算法的点云分割算法。在三维重建方面，分别对点云分割、3D特征提取和纹理映射等三维重建算法进行研究，提出了基于RANSAC算法的点云分割算法，并对一桌面场景进行实验验证。结果表明，对于背景是平面支撑的场景，该算法有很好的分割效果。 7、综合以上算法研究成果，开发了一个强辐射环境下基于图像和点云融合的软件系统。采用该系统对一桌面场景进行了空间目标的定位实验，得到了真实感较强的彩色三维点云模型，并利用图像与点云的空间映射关系，实现了空间目标的定位。 |
| Other Abstract | Intense radiation environments widely exist in nuclear power plant, civil irradiation plant, radioactive material storage and disposal sites as well as outer space, etc. Because intense radiation is harmful to human health, daily operation, maintenance, and accident emergency disposal and so on of the relation facilities in such places are mainly performed by robots or remote operation. Accurate terrain and location information of the worksite is the basis of robot or remote operation job, and this is mainly completed by vision system. However, the intense radiation can also damage the electronic equipments, higher requirements has been put forward for the establishment of the vision system. With the development of nuclear industry, aerospace and other industries, the relevant theoretical and technical studies of the vision system under intense radiation environment gradually become the most focused research topics in the field of modern science and technology. In this paper, with the production needs in areas such as nuclear power and nuclear irradiation, we aim at problems such as radiation protection, data processing, target location and information fusion of the vision system under intense radiation environment, and investigate related theory and key technology studies based on the fusion of image and point cloud data. The following results have been achieved: 1. An experimental platform for the hardware and supporting software of vision system under intense radiation environment were built. Based on analysis of the characteristics of existing vision system and the requirement of intense radiation environment, the visual mode of the system was determined to be the information fusion of camera and 3D laser range finder for their complementation. According to the task of the vision system under intense radiation environment, we give detail analysis to the hardware composition and configuration parameters. The configuration scheme of the system was given, and then a platform for the hardware system and supporting software was constructed. 2. Radiation damage mechanism of the vision system was studied and an effective radiation-resistant shielding device was developed. Based on the analysis of the research status and development trend of the resist radiation hardening method for the electronic system, the damage mechanism of intense radiation environment on the electronic system and the radiation-resistant properties of camera and laser range finder were studied. |
| Subject Area | 地球深部物质与流体作用地球化学 |
| Language | 中文 |
| Document Type | 学位论文 |
| Identifier | http://ir.gyig.ac.cn/handle/352002/5850 |
| Collection | 研究生_研究生_学位论文 |
| Recommended Citation GB/T 7714 | 刘成安. 强辐射环境下基于图像与点云融合的视觉系统研究[D]. 北京. 中国科学院研究生院,2013. |
| Files in This Item: | ||||||
| File Name/Size | DocType | Version | Access | License | ||
| 强辐射环境下基于图像与点云融合的视觉系统（4722KB） | 暂不开放 | License | Application Full Text | |||
Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.
Edit Comment