Automatic nozzle centering via NIR coaxial machine vision for high power laser cutting

What the paper says

Laser cutting has become a reference manufacturing technology for metal sheets and tubes due to its flexibility, productivity, and quality compared to conventional methods. Although widely adopted in industry, variations in process conditions, material properties, and calibration can affect cut quality. In particular, alignment between the laser beam and the nozzle center plays a key role in reducing defects such as dross attachment and surface roughness. This alignment is still often performed manually, leading to subjective and inaccurate results. An automatic nozzle centering system is therefore of great interest, as it reduces alignment time and errors by removing operator input and improving accuracy. In this research, an ad hoc centering system was investigated, employing a laser in a modulated manner, a coaxial camera monitoring system, and an external NIR illuminator. An image vision algorithm was implemented to detect the contours of the nozzle and the laser beam, enabling evaluation of misalignment error, while a feedback control scheme using two stepper motors was designed to correct it. Results show that the proposed solution can automatically aligning the laser beam with the nozzle center, achieving a residual error comparable to the manual method and enabling reduced human intervention and greater process adaptability.