EFPI Solution Type and Concentration Sensing System Based on ConvNeXt Network Model

To detect the type and concentration of the solution, an extrinsic fiber Fabry-Perot interferometer (EFPI) solution sensing system based on the ConvNeXt network model is designed in this article. In the experiment, we carried out three types of experimental samples, which were seven different solutions; six kinds of mixed solutions were obtained by pairwise mixing of seven kinds of solutions; five different concentrations of glucose solutions. After the flat end of the single-mode fiber is taken out from the solution, the droplet will form a fiber FP cavity, and the evaporation process of the solution will change the cavity length of the fiber FP cavity, which will cause a change in the reflected interference signal. To realize the detection of solution type and concentration, this article directly obtains the time domain feature set from the reflected interference signal, and then trains and tests the feature set based on the ConvNeXt network model. The experimental results show that we have achieved relatively accurate identification of the above three types of experimental samples, which proves that the identification of solution type and concentration based on the time-domain characteristic signal measured by EFPI is effective.