Research on Inversion Algorithm of Aerosol Extinction Coefficient Based on Elman Neural Network

Lidar, as an active remote sensing detection instrument, has become a powerful tool for atmospheric aerosol detection research. The extinction coefficient could be inverted by the lidar equation. However, the traditional method required many assumptions and complicated calculations when inverting the aerosol extinction coefficient, which greatly limited the accuracy and efficiency of the inversion. In this article, a method for predicting the aerosol extinction coefficient using Elman neural network was proposed. The neural network model was continuously trained to directly predict the aerosol extinction coefficient from the lidar echo signal, which effectively improved the aerosol extinction of the coefficient inversion efficiency. The experimental results show that the method with high prediction accuracy and the prediction effect was improved. The wide application prospect and practical value were possessed by the method and it provided a new idea for the inversion of extinction coefficient.