无损检验机器人焊缝识别技术研究

发布时间：2018-04-25 10:12

本文选题：无损检验机器人 + 焊缝识别　；参考：《深圳大学》2017年硕士论文

【摘要】：通常情况下,无损检验机器人的路径规划问题是通过人工标识或者位置要求严格的离线编程方式来解决的,而这两种路径规划的方式常常会受到检验环境的限制而导致检验效率低下、精度不高,同时也无法完全做到智能检验。所以研究焊缝识别技术在克服无损检验机器人传统路径规划的缺点、提升无损检验机器人的焊缝识别效率与精度、解决无损检验机器人的自动跟踪难题、实现焊缝无损检验的智能化与自动化等方面有重要的意义。在无损检验机器人焊缝识别技术研究中,焊缝识别方法研究是其关键。本文针对无损检验机器人的焊缝识别问题,研究不同的焊缝识别方法,并将研究的不同焊缝识别方法构成一套焊缝识别技术评价系统,同时建立了相应的算法性能评价指标。针对不同的焊缝工件,无损检验机器人焊缝识别技术评价系统根据算法性能评价指标选择一种合适的焊缝识别算法完成焊缝的识别。本文通过分析焊缝图像的视觉特征,提出了基于图像边界特征的焊缝识别算法;本文通过分析焊缝图像的统计特征,提出了基于图像纹理特征的焊缝识别算法。本文在焊缝识别算法研究中既含有普通的图像识别方法(如利用投影技术识别焊缝)又覆盖复杂的图像识别方法(如利用人工神经网络技术识别焊缝)。而且本文采用钢管焊缝与相贯线焊缝两种不同类型的焊缝对本文所提出的焊缝识别算法分别进行了仿真实验验证,实验验证的结果表明本文所提出的焊缝识别算法均可完成焊缝的识别。最后在Matlab平台上基于本文所研究的焊缝识别方法搭建了一个焊缝识别算法网络,构建无损检验机器人焊缝识别技术评价系统,并针对不同类型的焊缝分别进行了仿真实验,实验结果表明根据该评价系统中的性能评价指标可以从系统中选择合适的焊缝识别算法实现焊缝的识别。本文对无损检验机器人的焊缝识别技术进行了理论研究和实验测试,并根据提出的方法搭建实验平台,有助于提高无损检验机器人焊缝识别的精度与稳定性、解决无损检验机器人自动跟踪问题,为焊缝无损检验系统的智能性奠定理论基础。同时,本文所研究的内容也可为无损检验机器人的视觉引导技术研究提供理论指导。
[Abstract]:In general, the path planning problem of nondestructive testing robot is solved by manual marking or strict off-line programming. However, these two ways of path planning are often limited by the inspection environment, which leads to low efficiency and low accuracy, and it is also unable to achieve intelligent testing completely. Therefore, the research of weld seam identification technology in overcoming the shortcomings of traditional path planning of nondestructive inspection robot, improving the efficiency and precision of weld identification of non-destructive testing robot, solving the problem of automatic tracking of non-destructive testing robot. It is of great significance to realize the intelligence and automation of weld non-destructive testing. In the research of nondestructive inspection robot welding seam identification method is the key. In this paper, aiming at the problem of weld identification of nondestructive testing robot, different weld identification methods are studied, and a set of evaluation system of weld seam identification technology is constructed, and the corresponding algorithm performance evaluation index is established at the same time. For different weld parts, the evaluation system of nondestructive inspection robot weld identification technology selects a suitable weld identification algorithm according to the algorithm performance evaluation index to complete the weld seam identification. By analyzing the visual features of weld images, a weld recognition algorithm based on image boundary features is proposed, and a weld recognition algorithm based on image texture features is proposed by analyzing the statistical features of weld images. In this paper, there are not only common image recognition methods (such as using projection technology to identify weld seam) but also complicated image recognition methods (such as artificial neural network) in the research of weld seam recognition algorithm. In this paper, two different types of weld seam, steel pipe weld and intersecting line weld, are used to verify the proposed weld identification algorithm. The experimental results show that the proposed algorithm can be used to identify the weld. Finally, a weld recognition algorithm network is built on the Matlab platform based on the welding seam identification method studied in this paper, and a non-destructive inspection robot weld recognition technology evaluation system is constructed, and the simulation experiments are carried out for different types of weld seam respectively. The experimental results show that the proper weld recognition algorithm can be selected from the system according to the performance evaluation index of the evaluation system. In this paper, the welding seam recognition technology of the nondestructive testing robot is studied theoretically and experimentally, and the experimental platform is built according to the proposed method, which is helpful to improve the accuracy and stability of the weld seam identification of the nondestructive testing robot. The problem of automatic tracking of nondestructive testing robot is solved, which lays a theoretical foundation for the intelligence of weld nondestructive testing system. At the same time, the contents of this paper can also provide theoretical guidance for the study of visual guidance technology of NDT robots.
【学位授予单位】：深圳大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;TP242

【参考文献】