Risk assessment of valley debris flow based on the residual structure—Taking Nujiang Lisu autonomous prefecture as an example

Abstract

In order to evaluate the hazard of the debris flow in the valleys of Nujiang Prefecture, a neural network model that simultaneously recognizes digital elevation data and remote sensing data is proposed based on the residual structure. After the model is trained on the historical disaster data, it scores the hazard risk of debris flow in the remaining valleys. The test results show that the model’s hazard scores for typical valleys are basically in line with the results of the field investigation. The model can achieve an accuracy rate of 84% and a kappa coefficient of 0.81 on the valley classification task. The model can be used for risk assessment of debris flow valleys in the study area and has a certain reference value for the prevention and control of debris flow in the region.