An edge-computing-based hazard factor adaptive landslide early warning method

Abstract

The present invention discloses an edge-computation-based adaptive landslide early warning method for triggering factors, the steps comprising: acquiring historical displacement monitoring data samples of the landslide object to be monitored, and storing them on a server; constructing a candidate function library storing several landslide prior feature information functions, wherein the output of the landslide prior feature information functions is a landslide displacement prediction value, and transmitting the candidate function library of landslide prior feature information functions to the edge-computation computer via the Internet; when the landslide displacement prediction algorithm is initialized, randomly selecting several landslide prior feature information functions from the candidate function library to form the stack module of the algorithm; the edge-computation computer retrieves the historical displacement monitoring data samples from the server and trains the algorithm’s stack module to obtain a landslide displacement prediction model, which is stored in the edge-computation computer.