Deep Learning-Based Wildfire Image Detection and Classification Systems for Controlling Biomass

Table 1

Survey of forest fire image detection and classification.

Research	Title	Method	Accuracy	Classes	Year
Khan et al. [26]	Deep transfer learning for forest fire detection: A new dataset and performance benchmark	VGG19	95.0%	Fire and no-fire	2022
Park et al. [27]	Deep transfer learning for decision-supporting multilabel image classification in wildfire situations	VGG16, ResNet50, Densenet121	—	Fires, nonfires, flames, smoke, structures, and people	2021
Kaabi et al. [28]	In a video of a forest fire, the first smoke was found using a deep belief network	DBN	95%	Smoke and no smoke	2018
Sousa et al. [17]	Using enhanced datasets with transfer learning to detect fires	Inception-v3	93.6%	Fire and not fire	2019
Tang et al. [29]	Deep learning in a forest environment for image classification	ForestResNet	92%	Normal, smoke and fire	2021
Alexandrov et al. [18]	Deep learning in a forest environment for image classification	HaaR, LBP, YOLOv2, Faster R-CNN, SSD	87.4%, 81.3%, 98.3%, 95.9%, 81.1%	Smoke and forest background	2019
Li and Yu [30]	An efficient convolutional neural network algorithm for detecting fire	Yolo-edge	—	Nonforest fires and forest fires	2021