铝合金MIG焊熔池视觉信息检测及处理系统

【作者】 张晓雯；

【导师】 樊丁；

【作者基本信息】 兰州理工大学 ， 材料加工工程， 2004， 硕士

【摘要】 在材料和结构因素确定之后，焊接产品（或结构）的弧焊质量则完全依赖于焊接过程中熔池的熔化和凝固，因此熔池熔化状态的控制是弧焊质量控制的核心问题；该问题可以通过实时反馈控制熔池熔化状态和熔化位置这两项关键技术来解决，而解决这两项关键技术的前提是如何解决好熔池熔化状态和熔化位置的实时传感问题。近年来，针对碳钢脉冲GTAW过程的控制问题，焊接工作者进行了广泛而深入的研究，并在生产中得到初步应用。对于铝合金GTAW焊过程控制的研究还较少，铝合金脉冲GMAW焊过程控制的研究几乎为空白，这是因为铝合金表面反射全波段可见光，熔池与母材金属的灰度对比度很低，且MIG焊时电流较大，弧光强烈，普通的光学传感器无法获取铝合金焊接熔池区域的图像，导致控制过程失败。 本文针对铝合金脉冲GMAW焊的熔池视觉图像检测及处理系统，从以下几个方面进行了研究：首先采用图像传感的方法获得清晰的熔池图像；然后通过图像处理获得反映熔池形状特征的参数，为以后建模和焊接过程控制打下基础；最后采用主动光视觉传感技术，对焊接过程中的视觉焊缝跟踪技术进行了初步的研究。 被动式视觉传感方法是焊接过程传感的一种主要方法，其主要原理是利用电弧光照射熔池表面获得熔池图像。本文对这种传感方法在脉冲GMAW焊中的应用情况进行了深入分析，提出了采用由窄带、宽带滤光片，中性减光片和吸热片组成的复合滤光技术，从而解决了电弧光对成像的干扰问题，获得了清晰度高、对比度强的熔池图像。 为了描述熔池形状，本文采用了简单有效的特征参数，即熔池形状由熔池长度L_t、宽度W_t和半长比R_hl来确定。并且设计和开发了熔池图像处理算法，该算法由图像的灰值化、滤波处理、边缘增强、边缘检测及提取、几何特征参数提取等组成。通过对各种滤波方法以及边缘检测和提取算法的比较，最终采用了自适应维纳滤波处理和Canny边缘检测算子，经过实践证明此算法精度较高。 通过实验证明，所设计的系统可以对熔池图像进行处理，提取出熔池的最大宽度和最大长度。更多还原

【Abstract】 The quality of welded structure is fully dependent on the melting and solidifying process of weld pool after the material and structural parameters are determined. Therefore, controlling the melting state of weld pool is the key for the quality control of arc welding. This can be solved by the real-time feedback controlling both the pool melting state and the pool melting position in the welding process. Moreover the precondition of reaching this goal is the sensing melting state and pool melting position in real-tune. Recently, extensive and thorough research is carried out on control of carbon steel pulsed GTAW. However, research on aluminum pulsed GTAW were still little and the research on aluminum pulsed GMAW is nearly blank, because some factors such as thef surface of the aluminum can reflect the visible light in whole wave range and the contrast between the welding pool and the plate nearly is very weak lead. So it is difficult to take the image of weld pool of aluminum alloy by ordinary visual sensor.Aimed at inspect and process system for weld pool of aluminum pulsed GMAW, several aspects of the research is proposed in this paper. Firstly, clear image of weld pool is captured through visual sensing method. Secondly, shape parameters of weld pool are extracted by image processing, which is the base on modeling and controlling of welding process. Finally, primary research is carried out on seam tracking technology of welding process through active visual sensing method.Passive visual sensing method is the major method for sensing welding process, the principle of which is to capture weld pool image by the illumination of electric arc. In this paper, the application of passive visual sensing in pulsed GMAW is deeply analyzed. Through composite filter technology can solved interfered by electric arc and can achieved clear high contrast images of weld pool. For describing the shape of weld pool, this paper presented simple and valid shapeparameters for determining the shape of weld pool: the length lt, width Wt, and half-length ratioi Rhl of weld pool. At the same time, image processing calculus forextracting the shape parameters is designed and developed. And this calculus is composed by gray-scale processing, filter processing, edge enhancement processing, edge detect and extract processing. Compared with many methods of filter and edgedetectors, this paper finally adopts Wiener filter processing and Canny edge detector. Through practice proved this calculus has high precision.Through test proved that the system is designed can process image of weld pool, and can extract the max width and length of image of weld pool.更多还原