Drone imagery forest fire detection and classification using modified deep learning model

Abstract

With the progression of information technologies, unmanned aerial vehicles (UAV) or drones are more significant in remote monitoring the environment. One main application of UAV technology relevant to nature monitoring is monitoring wild animals. Among several natural disasters, Wildfires are one of the deadliest and cause damage to millions of hectares of forest lands or resources which threatens the lives of animals and people. Drones present novel features and convenience which include rapid deployment, adjustable and wider viewpoints, less human intervention, and high maneuverability. With the effective enforcement of deep learning in many applications, it is used in the domain of forest fire recognition for enhancing the accuracy of forest fire detection through extraction of deep semantic features from images. This article concentrates on the design of the drone imagery forest fire detection and classification using modified deep learning (DIFFDC-MDL) model. The presented DIFFDC-MDL model aims in the detection and classification of forest fire in drone imagery. To accomplish this, the presented DIFFDC-MDL model designs a modified MobileNet-v2 model to generate feature vectors. For forest fire classification, a simple recurrent unit model is applied in this study. In order to further improve the classification outcomes, shuffled frog leap algorithm is used. The simulation outcome analysis of the DIFFDC-MDL system was tested utilizing a database comprising fire and non-fire samples. The extensive comparison study referred that the improvements of the DIFFDC-MDL system over other recent algorithms.

Forest Defender Fusion System for Early Detection of Forest Fires

“…The Forest Defender Fusion System outperforms the other models with a precision of 99.86%. Mashraqi et al [9] and Khan et al [22] had a precision close to that of the system with precision scores of 99.38% and 97.42%, respectively, which highlights their efficacy for detecting forest fires. However, the precisions of 95.7% and 94.1% shown by Khan et al [23] and Sousa et al [24] were lower.…”

“…With an accuracy of 99.86%, the Forest Defender Fusion System outperforms all other models taken into consideration for this comparison. The model that is closest to the proposed system is the one developed by Mashraqi et al [9] with an accuracy of 99.38%. This implies that the Forest Defender Fusion System has a high detection accuracy.…”

“…The performance of the Forest Defender Fusion System was evaluated by comparing it to other fire detection models proposed by Khan et al [23], Mashraqi et al [9], Sousa et al [24], and Khan et al [22]. The Forest Defender Fusion System demonstrates a superior Recall value of 99.86%, outperforming the compared models (94.2%, 99.38%, 93.1%, and 99.47%).…”

“…Many studies have proved the success of using drones for the early detection of forest fires. Mashraqi, A. M. [9] proposed a model for the early detection of forest fires using RGB images taken from drones, while Wang, Y. [10] proposed a model using IR images.…”

Precision-Boosted Forest Fire Target Detection via Enhanced YOLOv8 Model

“…This served as inspiration for our work. Mashraqi Aisha M. et al [37] emphasized the design of forest fire detection and classification in drone imagery, using the modified deep learning (DIFFDC-MDL) model. Remarkably, they employed the Shuffled Frog Leaping algorithm and simulated the results using a database containing fire and non-fire samples, thereby enhancing the classification accuracy of the DIFFDC-MDL system.…”