Line laser measurement for aero-engine compressor blisk blades based on similarity-guided overlap registration

Line laser technology provides a flexible, efficient, and cost-effective measurement strategy for blade manufacturing. A critical step in this process is to stitch the local point clouds acquired from different views through coarse-to-fine registration. For new-generation aero-engines, compressor blades are evolving towards blisk structures, resulting in extremely low overlap between adjacent-view local point clouds. Moreover, the blades feature thin-walled free-form shapes, high-reflective surfaces, and variable curvature profiles, leading to captured local point clouds with insignificant features, noise, and inconsistent density. These factors, combined with the influence of blade machining errors, pose major challenges to fine registration. To address these problems, a similarity-guided overlap registration method, i.e., scale comparison and bidirectional interaction (SCBI), is proposed. This approach guides registration by leveraging the fact that overlap regions of adjacent-view point clouds are essentially identical. By enhancing scale and shape similarities in overlap regions, the robust registration of local point clouds is achieved. Additionally, a blade simulation dataset is developed to systematically evaluate six types of mainstream registration methods and SCBI, among which the latter achieves optimal results. Finally, a line laser-based on-machine measurement (LLOMM) system is established to measure compressor blisk blades, with a mean deviation less than 0.030 mm. Real measurement experiments are designed to elaborate its measurement deviation, machining error evaluation, and measurement efficiency. With detailed comparisons against four types of advanced measurement technologies, the LLOMM system demonstrates potential advantages in the above aspects. This study provides a new solution for blisk geometric error evaluation and its adaptive machining development.