Automated edge-feature-based high-precision alignment and splicing for multi-core fibers

High-precision fusion splicing is a prerequisite for the practical deployment of multi-core fiber (MCF) space-division multiplexing systems. However, the conventional manual alignment method suffers from low efficiency and a high risk of spatial geometry mismatch due to the complicated core arrangement of MCFs. In this article, we develop a high-precision and rapid automatic alignment method for MCFs based on a fiber end-face image processing technique. By extracting the edge features of the fiber cores and calculating their cross-correlation coefficient, the proposed method can automatically identify the most matched rotation angle, thereby ensuring accurate core-to-core matching and preventing geometric alignment errors. Crucially, a fusion splicer is controlled by a laptop with our self-developed software based on the proposed algorithm, which automatically retrieves MCF end-face images and executes alignment and splicing commands. The alignment process is completed in less than 20 seconds with an angular accuracy of 0.1°, and the total splicing time is less than 3 minutes. Furthermore, ultra-low average splicing losses of 0.068 dB (central core) and 0.123 dB (outer cores) are achieved for a 7-core fiber, and 0.079 dB for a 4-core fiber. These results verify the efficiency, reliability, and high quality of the developed automatic alignment and splicing technique for MCFs.