Application of three-dimensional point cloud technology based on spectral confocal in FPC welding inspection

Abstract In response to welding quality inspection challenges in new energy vehicle battery systems, a fusion detection method based on two-dimensional (2D) and three-dimensional (3D) data is proposed. Initially, the point cloud data of the region of interest (ROI) were extracted with point cloud percentile filtering. Subsequently, the processed point cloud data were projected onto a 2D plane, and the coordinate data of the ROI was obtained by virtue of 2D image-processing techniques. This data was back-projected into a 3D space to retrieve the corresponding depth data. Finally, the depth value of the weld pit was determined by computing the vertical distance between the bottom of the weld pit and a predefined reference plane from the obtained depth data. Experimental results indicate that when detecting workpieces containing 50 valid weld pits, compared with traditional 3D detection methods, the detection time of this method is reduced from 6.70 s to 1.18 s. Additionally, the depth measurement error and variance are reduced by 39.24% and 59.46%. This method significantly improves the detection efficiency and accuracy, enabling an efficient analysis of the welding quality.