An efficient 3D cutting scheme for detecting defects on products of complex geometry

The surface quality of parts is essential for the proper functioning of high-end equipment. Traditional 3D defect detection methods, based on geometric feature extraction, struggle with complex geometries, resulting in potential false or missed detections. This study proposes a surface quality defect detection scheme based on a 3D cutting algorithm. The inspection solution proposed in this work utilizes a 3D cutting algorithm to perform isometric multilayered cutting of a part. The results of each cross-section layer for defective and standard parts are aligned using an alignment algorithm for defect detection. Numerous numerical operations are converted into Boolean operations by reconstructing the 3D point cloud data with the proposed cutting algorithm. This scheme significantly reduces computing time and avoids the effects of complex surface geometries. Our defect detection solution has been validated not only on complex parts with holes but also on the manufacturing line of aeroengine impeller parts. The experimental results demonstrate that the defect detection rate is more than 95%, and the computational efficiency is increased 100x compared to traditional solutions. • Suitable for detecting defects on parts having complex surface geometry. • Conversion of numerical calculations to Boolean calculations for in-situ defect detection. • Outperform the traditional methods in terms of sensitivity & specificity of defect detection. • Up to 100x increase in computing efficiency over traditional solutions.