Surface Plasmon Resonance Sensor based on Ag Layer Coated PCF for Dry Sandy Soil Detection with Deep Learning Algorithm

This paper suggests an LSTM deep learning algorithm that can improve the sensing accuracy of a Photonic Crystal Fiber (PCF) based Surface Plasmon Resonance (SPR) sensor used in dry sandy soil detection. Optimized coupling efficiency between the plasmonic and core layers improves the sensor's refractive index sensitivity and detectable range. Training an LSTM model to predict resonance wavelength improves SPR sensor accuracy and real-time monitoring. The functionality of the sensor is simulated and analyzed using the Finite Element Method. In order to prevent oxidation, titanium oxide was applied to the visually appealing Ag plasmonic material. The sensor's 8 nm TiO 2 and 55 nm Ag improve resolution from 1.35 to 1.40 RIU. This sensor has high sensitivity at 1890 RIU ^-1 amplitude sensitivity and 19000 RIU ^-1 wavelength sensitivity, and experiments have shown that the sensor is sensitive to refractive index changes, has a large detectable range, and can accurately predict the resonance wavelength using deep learning. The proposed sensor design could reliably and cheaply detect dry sandy soil in agriculture and environmental monitoring.