In order to solve the problem of traffic congestion and unbalanced utilization of wavelength resources caused by the explosive growth of traffic demand in the next generation mobile communication network. In this paper, a multichannel all-optical wavelength converter with stable output power is designed by using germanium doped microstructure fiber. Based on the theoretical model of stimulated Raman scattering, a multi-channel all-optical wavelength converter is constructed. The waveforms of the pumped signal light and the converted detection light are analyzed, and the influence parameters of gain and flatness are discussed. The simulation results of OptiSystem show that the waveforms of the four-channel detection light and the pump signal are consistent and the output power is basically the same, and the gain fluctuation range is less than ±0.38dB, which verifies the feasibility of the method.
In order to better extract the infrared target information of images in dark scenes and retain more background texture details, an infrared and visible light image fusion algorithm based on fuzzy C-means clustering (FCM) and guided filter is proposed. Firstly, the target information is extracted from the source infrared image by FCM, and the target area and background area of the infrared image are obtained. Then, the target region coefficients and background region coefficients are decomposed into their respective high-frequency and low-frequency subband coefficients by using non-subsampled shearlet transform (NSST). Then, according to the different characteristics of different regions, different fusion strategies are adopted. In order to retain more target information, low-frequency subband coefficients of infrared image target area are selected as fusion coefficients of low-frequency target area, and high-frequency subband coefficients of infrared image target area are selected as fusion coefficients of high-frequency target area. In order to keep more texture details, the method of maximizing low-frequency subband image coefficients and information entropy is adopted in the fusion of low-frequency background region. The method of guided filter combined with dual-channel spiking cortical model (DCSCM) is used in the fusion of low-frequency background region. Finally, the final fusion image is obtained by NSST inverse transform. Simulation results show that compared with the existing algorithms, the fusion image obtained by this algorithm has prominent infrared target in subjective vision, clear background texture details and high hierarchy. In objective evaluation, the indexes are better than other algorithms as a whole.
We propose a method that uses the back propagation (BP) neural network algorithm to optimize the design of the multipump Raman fiber amplifier. We determine the optimal training model by examining the number of hidden layers in the multilayer BP neural network and the number of neural nodes contained in it. The model more accurately reflects the mapping relationship between the wavelength and output of the pump light and the Raman net gain distribution, instead of the traditional method of solving the Raman-coupled wave equation. The experimental results show that, using the trained BP neural network model to train new validation datasets, the studied Raman amplifier achieves the desired performance, and the maximum error between the target value and the predicted value does not exceed 0.3 dB. Compared with previous studies, this design scheme improves the accuracy of model calculation and the optimization efficiency of the Raman amplifier.