A deep learning-based scheme is demonstrated for the mode decomposition of Hermite-Gaussian(HG)beams.A convolutional neural network(CNN)is trained with simulated samples of multiple modes.To expedite the train⁃ing process,transfer learning is employed.The trained neural network can efficiently and accurately decompose HG beams across four modes using the intensity image of a single emission,providing both mode amplitude and phase infor⁃mation.Experimental results demonstrate that this method achieves high accuracy and stability in recognizing and de⁃composing various modes.This approach holds significant potential for applications in optical communications,optical imaging,laser processing,and other fields.
关键词
深度学习/Hermite-Gaussian光束/模式分解/迁移学习/模式识别
Key words
deep learning/Hermite-Gaussian beam/mode decomposition/transfer learning/mode recognition
维普
万方数据