大型锻件热态几何参数在线测量仪

Oriented to the urgent demand of high efficiency, energy saving and material reducing manufacturing for large forgings which are utilized in the core equipments of national major engineering projects in the fields such as nuclear power, ship building, chemical industry, national defense, energy power and iron and steel manufacturing, a real-time measurement technology for the geometric parameters of hot large forgings is developed based on machine vision method in this project. Firstly, the spectrum selective imaging model under variable temperature field for hot large forgings is proposed in order to acquire clear image characteristics of different typical large forgings. Then, the rapid recovery method of geometric parameters for hot large forgings is presented, based on multi-camera vision by virtue of laser aided labeling, and the seamless splicing for multi-camera images of hot large forging can be realized, so as to obtain whole geometric information of hot large forgings. Moreover, laser aided real-time multi-camera calibration method and calibration system are constructed to obtain fast and accurately cameras parameters in the workshop. Then, a set of comprehensive database for measurement process of hot large forgings is established for the engineering application. Finally, a set of real-time measurement instrument prototype for thermal geometric parameters of hot large forgings is developed based on multi-camera stereo vision and laser aided calibration, while application experiments of the measurement instrument are conducted in forging workshop such as Dalian Heavy Industry Group, to testify and promote the technical maturity, reliability and practicality of the measurement instrument. The research results can provide significant technical support for improving hot large forging material utilization ratio, manufacturing safety and production efficiency, making it possesses very wide application prospect.

面向核电、造船、化工、国防、能源动力、钢铁等领域中国家重大工程核心装备用大尺寸锻件高效、节能、节材制造的迫切需求，研究基于机器视觉法的大型锻件热态几何参数在线检测难题，提出时变温度场下大型锻件光谱选择成像模型，以获取不同典型大锻件材料的清晰图像特征；研究基于激光辅助标记多目视觉的大型锻件几何参数快速还原方法，实现大型锻件多目图像的无缝拼接，以获得大型锻件的整体几何信息；构建多目摄像机激光辅助在线标定方法和标定系统，实现摄像机锻压现场的快速精确标定；建立一套锻件测量工艺综合数据库，提高测量仪的工程实用性；开发出一套基于多目立体视觉及激光辅助测量的大型锻件热态几何参数在线测量仪样机，在大连重工起重集团等单位锻压现场进行大型锻件几何参数在线测量仪应用实验研究，验证并完善测量仪技术成熟度、可靠性和实用性。研究成果可为提高大型锻件材料利用率、生产安全性和生产效率提供重要的技术支撑，应用前景十分广阔。

面向航空航天、国防、能源动力等领域高端装备用关键零部件的高效、节材、节能锻造的迫切需求，准确、及时地对大型锻件的热态几何参数进行在线测量是改善锻造工艺、提高加工精度和生产效率的必要手段。目前，现有的锻件测量方法可分为接触式和非接触测量两类，人工接触式测量法具有测量效率低，精度差、操作人员安全难保证的问题，而现有的激光扫描、机器视觉等非接触测量方法存在测量效率低，环境适应性差等问题、难以满足在线测量的需求。因此，急需研究大型热态锻件在线测量关键技术与仪器，实现大型锻件三维几何参数的精确、快速、实时测量。.本项目针对大型高温锻件测量可靠性差、测量范围与测量精度难以兼顾的问题，提出了基于激光辅助的多目视觉测量方法。该方法的核心思想是利用投射到锻件表面的辅助激光光条强化锻件特征，通过激光光条间接表征锻件几何特征，结合多目视觉数据拼接实现锻件整体几何参数的准确、可靠、快速测量。针对现场强光辐射、环境光干扰导致锻件图像无法清晰获取的问题，提出了基于“光谱选择抑制高光、光强可控补偿成像”的热态锻件特征图像采集原理，形成了基于滤除高温辐射强光的光谱选择成像方法、时变温度下图像质量实时补偿方法，有效地解决了高温强光图像失真这一视觉测量的首要难题，实现了800℃-1250℃范围内清晰图像采集；针对锻件表面氧化皮脱落、烟雾粉尘等因素影响导致图像受噪声影响特征难以有效识别的问题，提出了“抑制背景噪声、强化光条特征”的图像处理思想，形成了基于信度评价的图像噪声抑制方法、光条特征对比度增强方法等，显著改善了图像清晰度；提出的基于多阶高斯中心提取、特征信息一致匹配的锻件尺寸测量方法，实现锻件特征尺寸快速准确测量；针对锻造现场设备布局受限、测量中受随机振动等因素干扰导致系统精度难以保证的问题，建立了视觉测量系统现场布局优化模型，提出了基于正交度补偿的视觉测量系统内参数离线标定和基于激光投影网格辅助的外参数快速标定及现场校准方法，校准后精度可达0.09%，解决了测量系统现场难标定、精度难保证的问题；最终研制出两台大型高温锻件热态几何参数在线测量仪样机，单次测量时间小于2秒，测量精度优于0.2%。.以实现大型高温锻件热态几何参数在线测量为目标，项目组研究出适用高温锻件在线视觉测量的图像获取、图像处理、精确提取、准确高效测量的新技术、新方法及测量仪器与软件，并在锻造现场得到初步应用。