一种基于图像质心的焊缝跟踪新方法

研究一种基于图像质心识别的电弧焊焊缝跟踪新方法。通过视觉传感器获取焊接区熔池图像，抽取图像质心坐标并构成状态向量，建立一种基于图像质心的状态方程和位置测量方程。在此基础上，应用卡尔曼滤波对图像质心位置进行状态估计，在时域中采取递推计算的方式得到最小均方差条件下的焊缝位置最佳预测值．从而消除过程噪声和测量噪声引起的焊缝位置测量偏差。计算机仿真和实际焊接试验结果显示该方法可有效地提高焊缝跟踪精度。     

熔池图像质心算法的焊缝位置测量模型

建立了一种基于熔池图像质心的焊缝位置测量模型,通过视觉传感器获取焊接区熔池图像,选择熔池前端为处理区域,对该区域进行中值滤波与图像灰度变换,并计算该区域的熔池图像质心值及相对应的焊缝偏差.在不同的焊接条件下,获取多组熔池图像及对应的样本数据,应用最小二乘法建立熔池图像质心与焊缝偏差之间的关系,得到基于熔池图像质心的焊缝位置测量数学表达式.在此基础上,通过分析比较各数学表达式之间的关系,建立焊缝位置测量数学模型.计算机仿真及焊接工艺试验结果表明,该模型可有效地检测焊缝位置.     

基于视觉图像传感的精密脉冲TIG焊焊缝跟踪

针对某复杂曲面薄壁不锈钢工件精密脉冲TIG焊中的焊缝跟踪问题 ,本文研究了一种基于视觉传感的高精度、高实时性焊缝跟踪技术。根据脉冲TIG焊的工艺特点 ,该技术通过选择特定波长的滤光片及合理的曝光时刻 ,采用工业CCD摄像机获取可直接分辨出焊缝、熔池和钨极的清晰、放大的焊接区图像。采用VC语言设计的图像处理算法 ,可以快速准确地识别出焊缝中心线 ,提取钨极偏离焊缝中心线的方向和距离 ,驱动步进电机调节焊炬位置 ,实现高精度、高实时性的焊缝跟踪。试验结果表明 ,该技术的单幅图像处理周期小于 12 0ms,能实现焊炬运动方向与焊缝偏差角小于 3 0°的焊缝跟踪 ,满足了复杂曲面的薄壁不锈钢工件的精密焊接要求。     

焊接视觉图像质心算法及其特性

从焊接视觉图像质心计算公式出发,根据图像灰度与视觉传感器响应电压值的关系、响应电压值与检测部位的发射率及光谱辐射强度等关系,由熔池特性,推导并建立了焊接视觉图像质心与焊接区各点温度间关系模型.通过焊接温度场的数值化,得到图像质心与焊缝中心关系模型的数值解及关系曲线,证明了视觉图像质心与焊缝中心存在相关性,为基于图像质心的焊缝跟踪算法提供理论依据.分析了质心算法的特性,为进一步研究实时焊接图像的质心及算法打下理论基础.     

磁光成像自适应卡尔曼滤波焊缝跟踪算法

焊缝跟踪是保证焊接质量的前提.针对0~0.05 mm的微间隙焊缝,研究一种色噪声环境下应用卡尔曼滤波实现焊缝跟踪的方法.通过对焊件施加磁场,利用法拉第磁旋光原理构成磁光传感器并获取焊缝磁光图像,提取焊缝中心位置构成状态向量,建立基于焊缝中心位置的系统状态方程与测量方程.针对系统过程噪声为色噪声,使用Sage自适应卡尔曼滤波,采用新息序列估计过程噪声协方差矩阵,准确预测焊缝中心位置.结果表明,根据自适应卡尔曼滤波方法能够有效提高焊缝跟踪精度.     

High speed fusion weld bead defects

Abstract

A comprehensive survey of high speed weld bead defects is presented with strong emphasis on the formation of humping and undercutting in autogenous and non-autogenous fusion welding processes. Blowhole and overlap weld defects are also discussed. Although experimental results from previous studies are informative, they do not always reveal the physical mechanisms responsible for the formation of these high speed weld bead defects. In addition, these experimental results do not reveal the complex relationships between welding process parameters and the onset of high speed weld bead defects. Various phenomenological models of humping and undercutting have been proposed that were based on observations of events in different regions within the weld pool or the final weld bead profile. The ability of these models to predict the onset of humping or undercutting has not been satisfactorily demonstrated. Furthermore, the proposed formation mechanisms of these high speed weld bead defects are still being questioned. Recent welding techniques and processes have, however, been shown to be very effective in suppressing humping and undercutting by slowing the backward flow of molten metal in the weld pool. This backward flow of molten weld metal may be the principal physical phenomenon responsible for the formation of humping and undercutting during high speed fusion welding.     

BP神经网络在焊缝位置识别中的应用

研究了一种焊缝位置识别新方法,在一定工艺条件下,使用视觉传感器采集焊接熔池图像,选取图像中熔池前端部分进行处理,先对其进行中值滤波与灰度变换,在此基础上,获取每一幅熔池图像的质心值、质心位移、质心速度及电弧与焊缝的偏差值作为训练样本数据.以质心值、质心位移和质心速度为输入量,以偏差值为输出量,利用BP神经网络建立其数学模型,最后对该模型进行检验.检验结果表明,该模型能够较准确地描述熔池图像质心与焊缝偏差之间的关系,为进一步实现精确的焊缝跟踪提供了理论和试验依据.     

电弧焊接焊缝偏差方法研究

焊缝跟踪技术是自动电弧焊接的一个重要研究领域,实现精确的焊缝跟踪对于提高焊接质量具有非常重要的作用。而要实现精确的焊缝跟踪,焊缝偏差(即焊缝中心与电弧的偏差)检测技术是一个关键。通过图像处理技术,选取熔池图像处理区域(包括熔池前端与熔池前端部份焊缝),并将熔池图像质心作为分析焊缝偏差的特性参量,研究利用熔池特性参数来建立焊缝偏差测量视觉模型的方法。     

铝合金TIG焊接熔池状态多传感器数据协同感知算法

针对铝合金钨极惰性气体保护电弧焊(tungsten inert gas arc welding,TIG焊)过程中,焊接工艺参数的实时状态与焊缝熔池三维尺寸间的非线性对应关系,研究建立一种基于信息物理融合的多传感器TIG焊过程熔池状态协同感知计算方法. 首先,构建由红外温度传感器、电弧形态传感器、电弧能量传感器和焊接位置传感器组成的TIG焊过程熔池状态信息物理融合系统架构. 其次,考虑焊接过程中焊枪电弧的运动特性和测量噪声影响,设计基于温度、位置、能量传感器信息交互的熔池长宽深三维参数状态感知策略,并基于多传感器数据的异步和异构特性,提出了基于无迹卡尔曼滤波的焊接过程中熔池状态的多传感器数据协同感知算法. 针对7075超硬铝合金TIG焊过程进行熔池参数在线测量与辨识试验,结果表明,所提算法能够根据TIG焊过程多传感器数据实时计算熔池参数结果,焊缝宽度和焊缝高度计算结果误差基本上控制在10%以内,该算法响应时间基本控制在0.3 s内,能够较为准确地评估焊接过程中熔池的实时状态.     

基于激光视觉的薄板焊缝跟踪方法研究北大核心

由于示教型焊接机器人在进行汽车薄板件连续焊工艺时存在装夹误差和热变形等问题,导致焊缝实际轨迹与示教轨迹存在较大误差。为提高焊接质量,基于焊接机器人构建激光视觉焊缝检测跟踪系统,提出基于目标估计准则的焊缝跟踪算法,实时跟踪焊缝中心点三维位置变化。以传统图像处理法提取初始帧焊缝特征点,通过改进的孪生神经网络对强干扰下的焊缝特征点进行跟踪提取。通过坐标转换得到机器人基坐标系下的焊缝中心特征点三维坐标。结果表明:该算法能精确提取跟踪焊缝特征点,平均误差为0.48 mm,平均帧率为90帧/s,优于传统图像处理方法和基于相关滤波的方法,能够实现快速准确的跟踪。     

焊缝偏差RBF神经网络预测算法

以10 kW大功率光纤激光焊接304奥氏体不锈钢板为试验对象,研究一种焊缝偏差预测算法.利用红外摄像机摄取焊接过程中的熔池红外图像,提取匙孔质心、匙孔形状参数和热堆积效应参数等反映激光束与焊缝位置偏差的特征量作为径向基函数RBF神经网络预测模型的输入量,建立焊缝偏差RBF神经网络预测模型.选择焊缝偏差特征量作为训练样本并对预测模型进行训练,建立焊缝偏差预测模型.结果表明,该模型能够对大功率光纤激光焊接过程中的激光束与焊缝位置之间的偏差进行有效预测.     

A machine vision approach to seam tracking in real-time in PAW of large-diameter stainless steel tube

Manual monitoring and seam tracking through watching weld pool images in real-time, by naked eyes or by industrial TV, are experience-depended, subjective, labor intensive, and sometimes biased. So it is necessary to realize the automation of computer-aided seam tracking. A PAW (plasma arc welding) seam tracking system was developed, which senses the molten pool and the seam in one frame by a vision sensor, and then detects the seam deviation to adjust the work piece motion adaptively to the seam position sensed by vision sensor. A novel molten pool area image-processing algorithm based on machine vision was proposed. The algorithm processes each image at the speed of 20 frames/second in real-time to extract three feature variables to get the seam deviation. It is proved experimentally that the algorithm is very fast and effective. Issues related to the algorithm are also discussed.     

基于卡尔曼滤波的焊缝偏差实时最优估计

建立了基于卡尔曼滤波的焊缝偏差实时最优估计算法.以焊缝中心位置为特征矢量,建立焊缝位置检测的状态方程和测量方程,并依据最小均方差原则建立了卡尔曼滤波最优估计的递推算法.测量噪声协方差由传感器测量误差的统计值得到,假定过程噪声是由于加速度变化引入,通过两点法确定焊缝中心位置的初值.在焊接过程中,应用卡尔曼滤波消除噪声干扰,实现焊缝位置的实时精确预测.计算机仿真和试验结果表明,焊缝偏差信号经过卡尔曼滤波处理后,消除了偶然因素和随机噪声的影响,提高了跟踪精度以及系统工作的稳定性,适合实际工程应用.

Abstract：

The optimal estimation algorithm for real-time welding deviation based on Kalman filtering is presented. The state equation and measurement equation for detecting the weld position is established, and the optimal estimation of the Kalman filtering recursive algorithm also is established according to the principle of minimum mean square error. Measurement noise covariance is obtained from the statistical value of measurement error, and after the process noise is supposed to derive from the changes in acceleration, the initial values of the welding center position are determined by the twopoint method. During the welding process, the welding position is accurately predicted while the noise interference is eliminated by Kalman filtering. The computer simulation and experiment results show that the weld deviation signal processed by the Kalman filtering can eliminate the disturbance of causal factors and random noise,improve the tracking precision and the stability of system, and be suitable for the practical engineering applications.     

多新息理论优化卡尔曼滤波焊缝在线识别

针对间隙小于0.05 mm的低碳钢对接焊缝,用磁光传感方法获取焊缝位置信息,研究多新息理论优化卡尔曼滤波在焊缝识别及跟踪中的应用.在获取磁光图像及提取焊缝位置的过程中存在较多干扰,而传统卡尔曼滤波受噪声的影响较大,难以对焊缝偏差进行最优估计.为此,结合多新息理论,提出一种焊缝位置检测的卡尔曼滤波改进算法,在对当前时刻进行预测时,充分考虑之前多个时刻的运动状态,综合历史数据估计出焊缝位置信息,对不同新息值进行试验比较并考虑计算量和滤波精度,发现选用两个新息值优化卡尔曼滤波算法可得到较好的效果.结果表明,多信息理论优化卡尔曼滤波算法可有效提高焊缝位置检测精度.     

Weld pool geometry during keyhole laser welding of thin steel sheets

Abstract

In industrial applications of laser welding it is often essential to obtain full penetration welds at high processing rates using minimal heat input. Keyhole welding meets these requirements when process parameters are kept close to the boundary where complete penetration switches to partial penetration welding. In the present work weld pool behaviour at the edge of the full penetration regime has been studied. Four types of keyhole penetration mode were observed. The first type is a completely developed keyhole through the material thickness and open in the root region, whereas the second type is closed at the root. The third mode is unstable and results in intermittent penetration involving periods of open and closed keyhole conditions interspersed with periods of lack of fusion. The fourth mode is a partial penetration mode. A possible explanation of the weld pool transient behaviour is presented based on three-dimensional reconstructions of the weld pools.     

三光路熔池图像的视觉注意特征提取

焊缝跟踪和熔透控制都是焊接质量控制中的重要部分.试验采用三光路视觉传感系统,在铝合金脉冲GTAW焊接过程中,三个光路同时采集熔池正前方、斜后方、斜下方图像,并投影到同一幅图像上.图像包含了焊缝、正面熔池和背面熔池等丰富的信息.采用视觉注意的方法找到图像各部分与熔池特征参数相关的小区域进行处理,将图像中包含的熔池特征参数提取出来.结果表明,视觉注意方法用于焊接过程中熔池特征的实时检测时,由于只处理感兴趣的小区域,比一般方法具有更高的明确性以及高效性.     

Misalignment detection and enabling of seam tracking for friction stir welding

Abstract

This paper describes a technique for determining the position of a friction stir welding (FSW) tool with respect to the weld seam during welding. Forces are used as a feedback signal, and a general regression neural network is trained to predict offset position given weld forces. Experimental results demonstrate the accuracy of the developed position predictor. This technique is proposed for online misalignment detection or as a position estimator for in-process tracking of the weld seam for FSW and robotic FSW.     

基于熔池图像尖端特征规律的焊接偏差测定方法

通过对管道全位置熔化极气体保护焊(gas metal arc welding,GMAW)熔池图像的研究分析,发现并验证了其根焊熔池图像尖端与焊缝坡口中心位置重合的现象,依据此规律性特征,提出了一种基于熔池图像尖端信息的焊接偏差测定方法.该方法的基本原理是,在对焊接过程熔池CCD图像进行中值滤波、小波变换、连通区域分割等图像处理后,以搜索算法测得的根焊熔池图像尖端位置信息作为焊缝坡口中心位置的坐标值,此值与焊丝中心坐标值之差即为焊接偏差量.试验证明,此方法能从根焊熔池图像中实时测定焊接偏差量,为实现机器人自动焊缝跟踪控制提供了可靠依据.     

Real time seam tracking system for automated fusion arc welding

Abstract

The demand for a simple and inexpensive yet flexible real time seam tracking system for automated fusion arc welding processes is ever increasing. To date, few techniques have been implemented in industry on a commercial basis, employing mostly thermovision devices, e.g. infrared vision cameras, infrared scanning radiometers, etc. or laser illuminated seam tracking systems. From the literature it appears that the use of simple vision sensors such as charge coupled device (CCD) cameras, which can monitor the surface temperature gradients as well as seam tracking operation in real time economically and reliably, has not yet been fully explored. It is thus conceived that further work is required in investigating real time seam tracking and adaptive control using image processing techniques and CCD cameras as they are inexpensive and robust as well as being sufficiently flexible. This highlights the importance of the present work, in which an attempt is made to extract geometrical parameters of the welding arc and to locate the weld seam ahead of the welding torch using images sensed by a simple CCD monochrome camera. The developed system has been tested successfully for fusion arc welding of butt joints with arc offsets, which are induced intentionally during experimentation.     

Weld shape variation and electrode protection under Ar–(Ar–O2) double shielded GTA welding

Abstract

Double shielded gas tungsten arc welding (GTA welding or TIG welding) of an SUS304 stainless steel with pure inert argon as the inner layer shielding and the Ar–O₂ active gas as the outer layer shielding is proposed in this study in order to investigate its effect on the tungsten electrode protection and the weld shape variation. The experimental results show that the inner inert argon gas can successfully prevent the outer layer active gas from contacting and oxidising the tungsten electrode during the welding process. The active gas, oxygen, in the outer layer shielding is decomposed in the arc and dissolves in the liquid pool, which effectively adjusts the active element, oxygen, content in the weld metal. When the weld metal oxygen content is over 70 ppm, the surface tension induced Marangoni convection changes from outward into inward, and the weld shape varies from a wide shallow one to a narrow deep one. The effect of the inner layer gas flowrate on the weld bead morphology and the weld shape is investigated systematically. The results showed that when the flowrate of the inner argon shielding gas is too low, the weld bead is easily oxidised and the weld shape is wide and shallow. A heavy continuous oxide layer on the liquid pool is a barrier to the liquid pool movement.