A method and apparatus for predicting atmospheric visibility

Abstract

An object of the present application is to provide a method and apparatus for predicting atmospheric visibility. The method includes acquiring target characteristic data, wherein the target characteristic data comprises upper air prediction data, surface prediction data and historical monitoring data; Extracting features from the target feature data to obtain sample data including basic features,spatial features and temporal features; Cutting the sample data according to a preset proportion to obtain training set data, verifying set data and test set data, and resampling the training set datato obtain resampling training set data; establishing A visibility prediction model according to the binary classification model, the regression model and the resampling training set data. The visibility prediction model is established through the processing of multiple feature extraction and pre-judgement results to obtain more accurate prediction results.

Description

technical field

[0001] The present application relates to the field of computers, in particular to a method and device for predicting atmospheric visibility. Background technique

[0002] Due to the complex meteorological environment factors, the index prediction of atmospheric visibility has always been a relatively complicated problem. At present, the commonly used prediction methods include the mechanism prediction method based on the atmospheric chemical transport model and the statistical prediction method based on the machine learning model. The former has been widely used in practical engineering. However, since the atmosphere is a very complex system, it is difficult to quantify it theoretically, so there are large errors in the mechanism prediction method.

[0003] At present, the forecast of atmospheric visibility by the domestic meteorological bureau is obtained by using the atmospheric chemical coupling model (WRF-Chem) combined with the correction of the multip...

Images