Differentiated Focal Plane Assisted Multi-core Fiber Fusion Splice Loss Evaluation

Multi-core fiber (MCF) is a high-capacity data transmission carrier based on spatial division multiplexing and is expected to be widely applied in internet data centers and submarine optical fiber cable transmission systems. However, the extensive use of MCF is restricted by the limited methods available to evaluate the MCF fusion splicing loss, which can profoundly affect the stability of the communication system. In this study, a scheme based on the differentiated focal plane method is proposed for MCF fusion splicing loss evaluation. The results revealed that the differentiated focal plane method can effectively extract fiber structural information of fiber cladding, center core, and outer core. Further, an accurate evaluation of a seven-core fiber fusion splicing loss is achieved based on the fiber structural deviation. A deep learning approach is used to improve the accuracy of loss evaluation to 99.1%. The proposed method can potentially handle more complex MCF and photonic crystal fiber.