With the continuous development of manufacturing, the requirements for product welding technology are becoming increasingly demanding. In order to improve welding quality and reduce the labor intensity and working environment challenges for workers, welding automation has become an important direction. However, many welding tasks have unique challenges, such as easy deformation of workpieces or inconsistent joints, which complicate automated welding. The Weld Navigator seam tracking system was specifically developed to address these issues. Let's take a look at the structure of the laser vision seam tracking system.
Laser Vision Tracking System Structure
The laser, acting as the structured light source, projects laser stripes onto the work surface below the sensor at a predetermined angle. The CCD directly observes the stripes on the surface under the sensor. The sensor is installed at a preset distance in front of the welding gun, enabling it to monitor the weld seam. During tracking, welding speed and forward viewing distance are used to calculate the delay time, ensuring that the welding gun follows the seam.
Laser Vision Sensor
The compact Weld Navigator laser vision sensor is typically installed on the side of the welding gun. The laser diode inside the sensor emits a visible laser beam, which is transformed into a fan-shaped light strip by the aperture, illuminating the workpiece. A CCD camera, set at an angle relative to the laser diode, captures the reflected light from the workpiece surface, forming an image of the joint's geometric profile on the CCD. Using optical triangulation, the system not only detects horizontal features but also determines the height between the sensor and the workpiece.
The sensor's function is to convert the measured physical quantity into a useful electrical signal that corresponds to the quantity, fulfilling the requirements for information processing and control. The control system analyzes the signal from the sensor and generates control signals. The actuator typically consists of a motor and cross slide, with the servo motor controlled by the control signals.
Image Preprocessing
In the welding robot vision system, image processing can generally be divided into two aspects: first, the enhancement, filtering, and binarization of the acquired raw images. The goal is to remove noise from the original image and make it clearer, with more prominent edge features—this is the preprocessing step. Second, the preprocessed image undergoes edge detection, feature extraction, and other operations to accurately obtain the center information of the weld seam for the robot's coordinate adjustments, ensuring welding quality.
In the study of real-time image processing methods for visual seam tracking, image enhancement is first applied to increase image contrast, followed by median filtering to remove image noise. Binarization is then used to extract the target image from the background. In post-processing research, key technologies for laser vision weld seam tracking system image processing, such as the extraction of the centerline of the weak light band and feature point detection, have been proposed with practical methods. The centerline extracted using the central axis transformation method is a continuous, single line. The slope analysis method is used to detect feature points in a convenient and reliable way. This processing method can accurately detect weld seam feature points, with fast processing speed, meeting the real-time requirements of the tracking system.
Operation of the Laser Vision Weld Seam Tracking System
The above content provides an analysis of the Weld Navigator laser vision seam tracking system's structure. Before the welding operation begins, the welding gun's position is adjusted to align with the center of the seam. Once the power is turned on, the system begins to operate. The program starts running, and the CCD in the sensor continuously monitors the image. The software triggers a timer, and based on the required welding speed, the time interval of the timer is adjusted. Every set interval, a snapshot of the image is taken, and the image is processed in real time by the program to find the weld seam's center. The deviation between the current seam center and the initial center is used as the detection value, and an adjustment value is sent to the robot controller. The controller then directs the robotic arm to guide the welding gun, achieving automatic tracking.
