Monitoring of high-power laser welding using high-speed photographing and image processing

•A visual system was used to monitor the plume and spatters during laser welding.•A new image processing algrithm of the feature extraction was proposed.•Experimental results could be used for laser welding quality inspection.

In conjunction with the rapid development of high quality laser welding, industrial manufacturing has seen a growing demand for process monitoring and diagnostics. This paper proposes an effective method for monitoring the high-power laser welding. By combining a high-speed camera with an ultraviolet and visible band-pass filter, high contrast images of a laser-induced plume and spatters were obtained at three different welding conditions. After preprocessing the images, the quantification of the plume and spatter features was performed. Plume detection was conducted based on the mechanism of a laser-induced plume. Using static and dynamic feature of spatters, a time delay tracking algorithm was designed for spatter detection. The results revealed that the features of plume and spatters, including plume size, plume growing direction, spatter radius, spatter ejected direction, spatter gray value and spatter velocity, were related to the welding quality. The proposed method could be used to monitor the stability of the laser welding process.