A prediction method of urban regional tail gas pollution based on depth spatio-temporal correction model

Abstract

The invention provides an urban area tail gas pollution prediction method based on a depth space-time correction model, which comprises the following steps: acquiring multi-source heterogeneous data;a self-encoder feature extraction method realizes feature dimension reduction extraction of the multi-source heterogeneous data by constructing a three-layer self-encoder network structure; tail gas emission correction, substituting dimension reduction characteristic data of each data source extracted in the step 2 into the tail gas emission correction model; Spatio-temporal series data generation; deep spatio-temporal network model pre-training; replacing the corrected model data with the telemetry data of the real monitoring points, and re-training the corrected regional tail gas emission prediction model; the weighted parameters of the model are determined to obtain a depth spatio-temporal network model, and the multi-source heterogeneous data are input to obtain a predicted regional tail gas pollution emission result.

Description

technical field [0001] The invention relates to the technical field of tail gas pollution prediction in urban areas, in particular to a method for predicting tail gas pollution in urban areas based on a deep spatio-temporal correction model. Background technique [0002] Urban pollutants are mainly produced by traffic emissions, the main pollutants are carbon monoxide CO, carbon dioxide CO 2 , Nitrogen oxide NO 2 Wait. CO is not only toxic, but also 2 Both are greenhouse gases, and the resulting greenhouse effect is a major hazard to the global environment. NO 2 It is the main substance that causes damage to lung function, so the prediction of exhaust pollution in urban areas is of great significance to environmental management and traffic planning. [0003] Existing methods for regional vehicle exhaust emission prediction can be roughly divided into two categories, namely classical diffusion models and satellite remote sensing. For classical diffusion models such as Ga...

AI Technical Summary

Problems solved by technology

Predicting vehicle exhaust emissions in each area of ​​the city faces the following four challenges: 1. Data sparsity and spatial heterogeneity; 2. Spatial dependence; 3. Time correlation; 4. External environmental factors

Images