Method for extracting surface pure microwave radiation anomaly before earthquake based on deep learning

Abstract

The invention discloses a method for extracting surface pure microwave radiation anomaly before an earthquake based on deep learning, and the method comprises the steps: building a mapping relation between multivariate parameters in data, such as surface temperature, surface vegetation index, surface soil humidity, rainfall and total cloud amount, in a historical earthquake-free period and surface microwave radiation through deep learning; and calculating an actual value of microwave radiation under the influence of non-seismic parameters in a pre-seismic target day by using the mapping relation, and finally obtaining pure seismic microwave radiation without the influence of non-seismic parameters through difference. According to the method, the potential influence of seismic activity on the surface temperature is fully considered, the mapping relation between the multivariate input remote sensing parameter and the surface temperature which are not easily influenced by the earthquake is established, the target daily non-seismic surface temperature is obtained, and then the surface non-seismic microwave radiation and the pure abnormal value caused by the earthquake are obtained accordingly; multi-element non-seismic parameter information of the target day is fully considered, but not only surface microwave radiation of the target day is considered; and the prediction rationality is ensured.

Images