DeepSmoke: Deep learning model for smoke detection and segmentation in outdoor environments

作者：

Highlights：

• Smoke detection and localization in both clear and hazy outdoor environments.

• Using a lightweight CNN architecture called EfficientNet for smoke detection.

• Employing DeepLabv3+ semantic segmentation architecture for smoke localization.

• Pixel-wise annotation of a new benchmark dataset for smoke semantic segmentation.

• Outperformed existing smoke detection and segmentation methods.

摘要

•Smoke detection and localization in both clear and hazy outdoor environments.•Using a lightweight CNN architecture called EfficientNet for smoke detection.•Employing DeepLabv3+ semantic segmentation architecture for smoke localization.•Pixel-wise annotation of a new benchmark dataset for smoke semantic segmentation.•Outperformed existing smoke detection and segmentation methods.

论文关键词：Smoke detection and segmentation,Semantic segmentation,Foggy surveillance environment,Wildfires,Disaster management

论文评审过程：Received 15 July 2020, Revised 4 December 2020, Accepted 25 April 2021, Available online 28 April 2021, Version of Record 7 June 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.115125