Wildfires are a crucial and urgent topic nowadays due to the significantly increased number of disasters caused predominantly by global warming. Li et al. (2022, p. 568) claim that technologies have reached certain progress and allow the protection of the people and environment with automatic fire detectors and alarms. Therefore, this problem statement aims to discuss the limitations of the intelligent fire management system and the consequences of these limitations.
The machine learning method works directly with the data and includes different tactics such as physics-based simulations so that firefighters can better understand the area they have to enter. However, Jain et al. (2020, p. 480) mention that one of the limitations is the possible inaccuracy and difficulty with the design because of the great number of regulations. The authors (2020, p. 497) also claim that one more possible limitation may include the inability to create an alternative in the observation when someone works with the data to predict the appearance of wildfire. Remote sensing is one of the intelligent fire management methods that can solve the problem. Molaudzi and Adelabu (2018, p. 222) mention several types of remotely sensed data such as Hyperspectral and Broadband. They have proved their efficiency in identifying the signs of a possible wildfire.
Nevertheless, despite their accessibility and effectiveness, those data types have several limitations, including the problem with the sensor saturation and pixels and experiencing biases. Moreover, the researchers state that “the absence of red-edge and narrow bands to target and highlight specific biophysical parameters” may complicate the process of fire prediction (Molaudzi and Adelabu, 2018, p. 231). Thus, the precision of the gained information might be confined to the inaccuracies in data.
Overall, the machine learning method and remote sensing are some of the most efficient practices of the intelligent fire management system. They allow for predicting the appearance of the fire to prepare for the disaster and, through the virtual design, study the area of the impact. However, both methods have their limitations, starting with certain inaccuracies in the data and biases regarding the area affected by the wildfire.
Jain, P., Coogan, S. C., Subramanian, S. G., Crowley, M., Taylor, S., & Flannigan, M. D. (2020). ‘A review of machine learning applications in wildfire science and management, Environmental Reviews, 28(4), pp. 478-505.
Li, X., del Río Saez, J. S., Ao, X., Yusuf, A., & Wang, D. Y. (2022). ‘Highly-sensitive fire alarm system based on cellulose paper with the low-temperature response and wireless signal conversion’, Chemical Engineering Journal, 431, pp. 567-571. doi: 10.1145/3436286.3436504
Molaudzi, O. D., & Adelabu, S. A. (2018). ‘Review of the use of remote sensing for monitoring wildfire risk conditions to support fire risk assessment in protected areas’, South African Journal of Geomatics, 7(3), pp. 222-242.