A High-Precision Workpiece Positioning System Based on Super-Resolution Microscopic Vision in Ultra-Precision Machining

Ultra-precision machining generally involves multiple process steps, during which a workpiece needs to be measured offline and re-mounted. Therefore, re-clamping errors apt to cause misalignment of the workpiece between machining stages are inevitable. Even small in-plane rotations and translations can significantly degrade feature alignment and final form accuracy of the machined components during multi-step processes. To address this issue, this paper proposes a high-precision on-machine positioning system based on super-resolution microscopic vision towards accurate measurement of re-clamping errors. Specifically, the system directly uses inherent micro-scale features on the machined surface as intrinsic references to guarantee positioning accuracy. In-situ microscopic images are captured before and after re-clamping operation, and then a coarse-to-fine registration pipeline is employed. During the coarse positioning stage, robust geometric feature extraction and geometry constrained fitting can provide a reliable initial three-degree-of-freedom (3-DOF) pose estimate over large rotation and translation ranges. During the fine positioning stage, 3-DOF sub-pixel template matching is performed to accurately estimate the residual positioning error within a small search range. Furthermore, to overcome the resolution limitations of conventional microscopy, a super-resolution reconstruction module tailored to microscopic imaging is used to enhance the effective resolution of microstructure images during fine positioning. Systematic experiments demonstrate that the proposed system can achieve arcsecond-level rotational accuracy over a rotational range of 180° and translational accuracy on the scale of tens of nanometers over a translational range of 180 μm. This proposed approach is a non-contact, high-accuracy, cost-effective, and readily deployable method for practical machining environments, and it has exhibited strong potential for re-clamping pose identification and high-precision workpiece reposition during ultra-precision machining operations.