Fire detection method and system

The invention discloses a fire behavior detection method and system, an improved YOLOv5s target detection algorithm is adopted to detect a fire behavior, and an improved CA attention module is added in a YOLOv5s network, so that the detection network can better position and identify a target, the ac...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: ZHAO HUAN, FAN WENTAO, LIU YUANMAN, HUANG NAN, MA JINGRUI, ZHAO YING
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 a fire behavior detection method and system, an improved YOLOv5s target detection algorithm is adopted to detect a fire behavior, and an improved CA attention module is added in a YOLOv5s network, so that the detection network can better position and identify a target, the accuracy of a target position is improved, and a map system can accurately position the fire behavior position later. The improved CA attention module is light enough, does not increase more calculation amount, does not affect the detection speed, and focuses attention on the texture features of flame and smoke. In order to improve the detection capability of a small target, a small target enhancement module is added to the neck network, a shallow feature map is used to compensate for scarce small target information in a deep feature map, and meanwhile, when an improved YOLOv5s network model is trained, a VariFocal Loss loss function is introduced, so that the detection capability of the small target is improved, and