Emergency event grading and classification method and device based on deep convolutional neural network

Abstract

The invention relates to an emergency event grading and classification method and device based on a deep convolutional neural network. The method comprises the steps of acquiring image information ofeach category of pre-classified emergency events, and building an emergency event image library; performing importance marking on each image in the image library; selecting a sample set from the marked emergency event image library; building a deep convolutional neural network model, wherein input of the model is an emergency event original image, and output of the model is an emergency event grading and classification vector; inputting the sample set into the model to perform training, and selecting a model by using S-fold cross validation; inputting an image of an emergency event to be graded and classified into the model for calculation, and obtaining a classification and grading result of the emergency event to be graded and classified. According to the method, the deep convolutional neural network is trained through historical image data so as to automatically learn features of an emergency event, thereby performing grading and classification on the event, and assisting a decisionmaker to master conditions of the emergency event scientifically and objectively.

Description

technical field [0001] The invention relates to the technical field of smart cities, in particular to a method and device for classifying emergency events based on a deep convolutional neural network. Background technique [0002] At present, there are usually two forms of grading and classification of emergencies at home and abroad: one is purely manual judgment. According to the situation of historical emergencies, the relevant core features are manually summarized to form an index system. Indicators, manually judge the type and level of the event; the second is manual + automatic judgment, first summarize the core characteristics of the emergency manually, and form an index system. When a new emergency comes, the machine will calculate the type and level of the event. level. [0003] Existing methods for dealing with emergency classification and classification all use traditional machine learning techniques, such as Bayesian networks, SVM support vector machine algorithm...

AI Technical Summary

Problems solved by technology

However, the selection of emergency features by these methods is done manually, and there is a certain degree of subjectivity and arbitrariness.

Images