摘要：窄间隙焊接技术是一项在各个工程领域受到广泛应用的焊接技术，随着时代的发展，工程上对窄间隙焊接的应用效率提出了更高的要求。传统的人工窄间隙焊接操作存在人为因素的影响大，焊接质量稳定性低，工作效率低等无法避免的劣势，很难满足高效实现的要求，所以实现窄间隙焊自动化就成为了提高效率的最有效的解决方案。窄间隙焊接自动化系统要能够稳定高效运行，必须建立能够实时准确的识别窄间隙焊坡口形状的识别系统，保证控制信息的准确。

本文首先建立了基于视觉传感信息的坡口识别实验系统，包含视觉传感模块，机械夹持传动模块以及计算机图像处理模块。三个模块协同合作能够良好的运行实验流程。在搭建好视觉实验系统的基础上，根据机器视觉的原理与各个实验设备之间的相对位置关系，通过数学推导建立了以视觉传感信息的像素坐标为参数的，刻画窄间隙焊坡口表面形貌的数学模型。

本文结合实际实验情景，对所建立的视觉系统进行视觉参数的标定求解，同时提出了处理传感图像的图像智能处理算法，此算法封装了自适应的阈值分割算法、面积滤波去除噪声算法与基于行列扫描的特征线段提取算法这三个子算法，能够以不高的硬件要求快速的运行求解程序得出坡口图像的特征点，同时与其他图像处理算法相比有着更好的处理效果。最后根据模型求解结果得出窄间隙坡口形状，与实际误差值在0.6%左右，结果精确。

关键词：窄间隙焊；坡口模型；机器视觉；特征提取算法

Abstract：Narrow gap welding technology is widely used in various engineering fields of welding technology. With the development of the times, the project on the narrow gap welding application efficiency put forward higher requirements. The traditional manual narrow gap welding operation has the influence of human factors, the stability of welding quality is low, the efficiency is low and so can not be avoided. It is difficult to meet the requirement of high efficiency. Therefore, the realization of narrow gap welding becomes the most efficient Effective solution. Narrow gap welding automation system to be able to stable and efficient operation, must be established to real-time accurate identification of narrow gap weld groove shape recognition system to ensure the accuracy of control information.

In this paper, a groove recognition experiment system based on visual sensing information is established, which includes visual sensor module, mechanical clamping transmission module and computer image processing module. Three modules work together to run well . Through the mathematical derivation, the mathematical model of the groove surface is characterized by the pixel coordinates of the visual information. 

Based on the actual experimental scenario, the visual parameters of the visual system are calibrated and the algorithm of image processing is put forward. The algorithm encapsulates the adaptive threshold segmentation algorithm, the area filter removes the noise algorithm and Based on the algorithm of the line segment extraction algorithm, the three algorithms can obtain the feature points of the bevel image with the fast running of the program without the high hardware requirements, and have better processing effect compared with other image processing algorithms. Finally, according to the model results, the narrow gap groove shape is obtained, and the actual error value is about 0.6%. The result is accurate.

Key words: narrow gap welding; groove model; machine vision; feature extraction algorithm

目录

第一章绪论 1

1.1课题的背景及其意义 1

1.2窄间隙焊的发展 基于视觉的窄间隙焊坡口识别+源代码:http://www.youerw.com/jisuanji/lunwen_203847.html