Identification and Classification of Fugitive Dust Sources in the Urban Construction Based on U-Net Model Structure with Remote Sensing Images
Various types of construction dust inevitably accompany the different phases of urban construction.Among them,the demolition areas and construction sites in the early,middle and late phases of construction are four important fugitive dust sources generated by urban construction.Taking the plain area of Ningbo City as the research object,based on the semantic segmentation of Google Earth high-resolution remote sensing image between March and August in 2019 with U-Net model,this research studied the identification and classification of fugitive dust sources.It found that the model could assist in the identification and classification of four types of fugitive dust sources,and the mixed correction strategy combining the artificial identification and spatial relations could be uesd to correct the results of model classification in local areas that had a low efficiency in identifying and classifying the dust types.And 940 sites of fugitive dust sources covering an area of 4430.23 ha were identified.It concluded that the sources of these four dust types were mainly concentrated in the middle part of the central city,and there were more fugitive dust sources in the demolition areas and in the early phase of construction,while there were relatively few fugitive dust sources in the middle and late phases.
NETL
NSTL
万方数据
