Dual-Output Mode Analysis of Multimode Laguerre-Gaussian Beams via Deep Learning

Abstract

The Laguerre-Gaussian (LG) beam demonstrates great potential for optical communication due to its orthogonality between different eigenstates, and has gained increased research interest in recent years. Here, we propose a dual-output mode analysis method based on deep learning that can accurately obtain both the mode weight and phase information of multimode LG beams. We reconstruct the LG beams based on the result predicted by the convolutional neural network. It shows that the correlation coefficient values after reconstruction are above 0.9999, and the mean absolute error (MAE) of the mode weights and phases are about 1.4 × 10-3 and 2.9 × 10-3, respectively. The model still maintains relatively accurate prediction for the associated unknown data set and the noise-disturbed samples. In addition, the computation time of the model for a single test sample takes only 0.975 ms on average. These results show that our method has good abilities of generalization and robustness and allows for nearly real-time modal analysis.

Measurements of plasma density profile evolutions with a channel-guided laser

“…The CCD is synchronously triggered by the DG645 device and is able to acquire the transverse, time-integrated intensity profile of the laser pulses. The quasi-Newton algorithm [97] and the model identification technique [98] can be studied to fit the intensity images with Equation (15). The two rows of subfigures are respectively the experiential profiles (lower ones) and the reconstructed profiles via fitting results (upper ones) in Figure 5, where the intensity is normalized.…”

The measurement of complex modal coefficients of a superposed vortex beam based on intensity sampling methods

“…for example, Laguerre-Gaussian (LG) basis modes, is demultiplexed by a modal decomposition method. The decomposition process has led to extracting information about the laser beam, such as the OAM of light [12], wavefront and quality factor [13], and the modal coefficients [14].…”

“…The measurement error for this technique did not exceed 3%-4% for amplitudes and 5%-6% for phases. Dual-output mode analysis method [14], and the convolutional neural network method [23] are the other two techniques based on deep learning, that in simulation can accurately calculate both the modal weight and phase information of a superposition. However, a complete comparison between the whole area of beam profiles is required to perform modal decomposition.…”