Remote sensing image road extraction method based on attention mechanism improvement

The invention discloses an attention mechanism improvement-based remote sensing image road extraction method. The method comprises the following steps of preparing a data set; building a remote sensing image road extraction network; and training the remote sensing image road extraction network. Acco...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: WANG XINRUI, LI PINRU, WEI DEBIN, XU YONGQIANG, WEN JINGLONG, XIE HONGJI
Format: Patent
Sprache:chi ; eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:The invention discloses an attention mechanism improvement-based remote sensing image road extraction method. The method comprises the following steps of preparing a data set; building a remote sensing image road extraction network; and training the remote sensing image road extraction network. According to the method, the pre-trained DenseNet-121 is adopted as an encoder of the remote sensing image road extraction network, and the pre-trained weight on a large-scale image data set is fully utilized, so that the road extraction network can better understand and capture key features in road images; the performance and accuracy of a road extraction network in a road extraction task are greatly improved, and a reliable technical basis is provided for research and application in related fields. According to the invention, a cyclic cross attention module and a convolution attention module are introduced into the road extraction network, and the overall structure and characteristics of the road can be better unders