Capsule neural network and adapted golden search optimizer based forest fire and smoke detection

Area of Science:

• Environmental Science
• Computer Science
• Artificial Intelligence

Background:

• Forest fires pose significant risks to ecosystems and human health, with increasing frequency exacerbating global warming.
• Accurate and timely detection of forest fires and smoke is crucial for effective mitigation and response efforts.

Purpose of the Study:

• To develop an innovative methodology for detecting forest fires and smoke using an enhanced capsule neural network (CNN) and an adapted golden search optimizer (AGSO).
• To improve the accuracy and dependability of automatic forest fire detection systems.

Main Methods:

• Utilized an enhanced capsule neural network (CNN) integrated with an adapted golden search optimizer (AGSO).
• Employed advanced deep learning and optimization strategies to identify complex patterns associated with wildfires.
• Tested the model on wildfire smoke imagery and the BowFire dataset.

Main Results:

• The proposed methodology demonstrated superior performance compared to traditional feature selection and classification methods.
• The integration of the modified CNN and AGSO significantly enhanced the accuracy and dependability of forest fire identification.
• The system facilitated rapid response and mitigation efforts.

Conclusions:

• Advanced computational techniques, such as combining capsule neural networks with golden search optimizers, are vital for tackling complex environmental issues like forest fires.
• The developed methodology shows great potential for progressing automatic forest fire detection systems, reducing risks, and ensuring safety.