Inversion of middle atmosphere temperature with Rayleigh Lidar based on BP neural network

Middle atmosphere temperature variation is an important monitoring parameter of earth environment evolution. Its observation is limited to remote sensing technology, among which active LiDAR is one of the limited local conventional observation technology. However, the impact of system noise and aerosol on the reliability of observation data has not been effectively eliminated. In order to obtain a more accurate middle atmosphere temperature profile through Lidar technology, a new method of middle atmosphere temperature inversion by Rayleigh Lidar based on BPNN was proposed. Based on BPNN, the mathematical relationship between atmospheric density and atmospheric temperature was analyzed. The network model is improved. In order to verify the validity of the model, the evaluation indexes are listed and calculated. The experimental results show that the improved BPNN can effectively improve the influence of gas system noise and aerosol on LiDAR temperature inversion when the double hidden layer is used.