Zhiguo Yan

A fast and robust method for line detection based on image pyramid and Hough transform

It is of great importance to locate lines quickly and robustly in images. In this paper, a new method based on a binary image pyramid and Hough transform is developed to extract lines fast and accurately in two steps. Lines are quickly detected via a two-dimensional Hough transform in a lower-resolution image on a pyramids top, and then they are finely located via a one-dimensional Hough transform in the original image at the pyramids bottom. The proposed method has good ability to detect lines in noisy images, with reduced computation cost and storage requirement. The performance analysis and experiments are provided to verify the effectiveness of the proposed method.

Recent developments on welding image processing and features extraction

www.instmc.org.uk www.instmc.org.uk Measurement + Control Vol 40/5 June 2007 • 139 R obots have been introduced to welding processes by their high accuracy, repeatability and reliability, which improve the productivity of welding industry greatly [Wu1]. Nowadays the welding industry is developing into multiple varieties and middle or small batch production modes, which require welding robots to be able to operate within the welding environment with more adaptability and intelligence. Welding robots guided by sensors can adjust the track, pose of torches and parameters automatically in welding process. Seam tracking is one of the principal problems in welding automation with high quality[Wilson2, Kuo3]. Vision control plays a more and more important role in robotic welding. With the development of machine vision in the past few decades, the technology of visual measurement and location can be used in many tasks such as robotic assembly, tracking and navigating. In the robotic welding field, visual measurement can provide relative position information for the visual control of welding robots. Structured light vision provides abundant information about work pieces without contact, and this method is widely used as one of the most promising sensing modes because of its accuracy, reliability, high signal-to-noise ratio and good performance in realtime [Cederberg4, Li5, Haug6]. Vision information in structured light images is represented by the laser stripes which include scale, shape and position of weld joints on weldment [Sicard7]. The tasks of image processing and features extraction are weld joint recognition and positioning. In the process of welding, structured light images suffer badly from strong arc light, splash and fake imaging with optical intensive glistening in the weld joints area caused by overdriven reflection on polished surface of weldment. These can give rise to difficulty, error and even failure in processing of the welding seam images and features extraction. In order to effectively eliminate the effects of disturbances in structured light images, some intelligent algorithms, such as fuzzy logic, template matching and neural networks etc, have been developed for the robust recognition and extraction of the weld joint. However, the disturbances and complexity of images in welding processes are still challenging issues to be addressed for the image processing algorithms. In this paper, some algorithms are presented on basis of the character analysis of laser stripes and disturbances in images under different welding conditions.

A Visual Servoing System for the Torch Alignment to Initial Welding Position

It is an important aspect to locate the initial welding position automatically. In this paper, a visual servoing system is proposed. Two CCD cameras are employed as the senseors to get the information of the welding seam. The designed visual servoing controller, which is based on Takigi-Sugeno (T-S) fuzzy model, is applied to the stepper motors to adjust the welding torch in the up-down and left-right direction. Experimental results are provided to demonstrate the validity of the proposed system.

Girth Seam Tracking System Based on Vision for Pipe Welding Robot

A girth seam tracking system based on structured light vision for pipe welding robot is presented. Firstly, the vision sensor is designed based on the analysis of laser reflection on weldment surface and the position relation of camera, laser plane and weldment for the effect of laser stripes imaging. Image processing and features extraction algorithms are developed in case of strong multiple reflections disturbance in welding seam image. An image-based vision control system of welding robot is adopted for girth seam tracking of pipe weldment. Girth seam tracking experiment by pipe welding robot was conducted to verify the performance of the system

Vision tracking system for narrow butt seams with CO2 gas shielded arc welding

An image-based visual system is proposed for narrow butt welding seam tracking in container manufacturing. A smart camera is used as the sensor to measure the welding seam. The feature extraction algorithm is designed based on the smart camera, which is used to compute the error between the welding torch and the welding seam. The reference feature is determined with many frames of images based on their statistical property. The current estimated feature is checked based on the reference feature. The abnormal features caused by strong noises such as welding arc light, splashes and fog are rejected. The normal features are used for welding seam tracking. A compact vision system for tracking seams is presented, which employs a programmable controller to control a stepper motor to eliminate the error detected by the smart camera. Experiments are well conducted to verify the effectiveness of the proposed system and method.

A Survey of the Sensing and Control Techniques for Robotic Arc Welding

www.instmc.org.uk www.instmc.org.uk Measurement + Control Vol 40/5 June 2007 • 147 A rc welding plays an important role in modern manufacturing. Although the welding techniques have been much developed in last decades, most of the welding robots serving in practical production are still with the mode of teaching and playback and have little adaptive ability to circumstance. Therefore, they can hardly meet the diverse requirements of users. Recently, many researchers have put their efforts on the automation of the welding process to improve the welding efficiency. This leads to the opening of a very challenging area of research, which concerns with robotics, sensor technology, artificial intelligence and some other relevant fields. A welding automation system should have two essential functions. The first function is the ability to control the position and orientation of welding torch in order to track the weld seam with a high precision. The second function is the adjustment capacity of the welding process variables, such as welding current, voltage, torch speed, wire feed rate and etc, in real time. Sensing and control techniques are the two most important keys to achieve these two functions. The welding features, such as weld joint dimensions and the welding pool geometry, should be detected accurately in the control of welding process. This means that the sensing technology is required. Up to now, many sensing strategies have been exploited in this field. Control strategies are also important for welding automation. Proper control method should be adopted to tune the welding process online so as to obtain a decent welding result according to the welding features measured by sensors. Generally, traditional control strategies are not sufficient to achieve a good performance due to the complex and nonlinear characteristics of welding process. Fortunately, intelligent techniques, such as artificial neural network and fuzzy logic, provide with a new way to conduct the control task. The motivation of this paper is to investigate the sensing and control techniques in robotic arc welding by presenting an extensive literature review. The rest of this paper is arranged as follows. In section 2, the sensing and control methods for seam tracking are discussed. Section 3 presents the ways to detect the weld pool, where the control strategies for welding process variables are also discussed. At last, a brief conclusion is presented in section 4. 2 Weld seam detection and tracking 2.1 Weld seam detection

A Vision-Based Seam Tracking System for Submerged Arc Welding

Welding seam tracking is a key issue to welding automation. In this paper, a seam tracking system based on laser structured light is presented. The paper discusses the structure of this system, and also explains the system’s working principle especially the image processing algorithm. The outside-welding experiments of steel pipe show that this system is very robust to various welding noises and it can improve the welding efficiency greatly

Design and Coordinated Motion Control of a Welding Robot for Large Scale Workpieces

In order to meet the demands of the welding of large scaled workpieces, a welding robot is designed with 9 motion joints and 6 degrees of freedom. Two groups of macro and micro motion mechanisms are adopted for large travel and high positioning precision. The welding trajectory is taught grossly in Cartesian space, and the motion of joints is planned in joint space with the kinematics model of the welding robot. A method of motion control of macro and micro joints is presented for stability and precision in seam tracking. Experimental results verified the effectiveness of the mechanisms of welding robot and the control system.