Robotic Fabric Alignment System for Sewing Using Global Local Weighted ICP
This work addresses the challenge of simultaneously aligning fabric edges and stitching lines with high precision prior to sewing, particularly under occlusion and arbitrary initial poses. To this end, the authors propose an automatic alignment system based on a Global-Local Weighted Iterative Closest Point (GLW-ICP) algorithm. The method integrates a weighted registration strategy that combines global edge points and local stitching-line points, while dynamically rejecting outlier correspondences from occluded regions during point cloud registration, thereby significantly enhancing both robustness and accuracy. Coupled with CAD model matching and robotic manipulation, the system achieves millimeter-level alignment accuracy across diverse fabric shapes, demonstrating its effectiveness and reliability in real-world sewing scenarios.