Scholarlink

With the aid of machine learning and Internet of Things technology, this current paper presents a cutting-edge approach to the detection of forest fires. Raspberry Pi 4 serves as the processing module. Inefficiency and delay cause the conventional detection mechanisms to barely detect fire until massive damage has occurred. Our system used temperature, humidity, smoke, and flame sensors and maintained visual information by way of a Raspberry Pi camera. For accurate and real-time fire detection, Convolutional Neural Networks (CNNs) are utilized for image processing of such visual input. The messages are notified through Telegram, and an Android application designed using MIT App Inventor also displays the current status. Other than this, K-Nearest Neighbor (KNN) and Artificial Neural Networks (ANN) models were also designed and compared based on their efficiencies. The CNN model performed the fire detection tests more effectively, and it surpassed the others in recall and precision. It is the most suitable system to be deployed in remote forest regions due to its scalability, affordability, and efficiency. Active conservation and fire management practices are empowered by its real-time performance. Its AI and IoT synergy make it revolutionary in disaster prevention, with response time sped up and lowered monetary and environmental expenditure.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Forest Fire Detection System Using IoT and Convolutional Neural Network

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Ashok Battula,
  • E. Mina,
  • Vamshi Gajjala,
  • Radha Krishna Karne,
  • Pradeep Reddy Kumbala

摘要

With the aid of machine learning and Internet of Things technology, this current paper presents a cutting-edge approach to the detection of forest fires. Raspberry Pi 4 serves as the processing module. Inefficiency and delay cause the conventional detection mechanisms to barely detect fire until massive damage has occurred. Our system used temperature, humidity, smoke, and flame sensors and maintained visual information by way of a Raspberry Pi camera. For accurate and real-time fire detection, Convolutional Neural Networks (CNNs) are utilized for image processing of such visual input. The messages are notified through Telegram, and an Android application designed using MIT App Inventor also displays the current status. Other than this, K-Nearest Neighbor (KNN) and Artificial Neural Networks (ANN) models were also designed and compared based on their efficiencies. The CNN model performed the fire detection tests more effectively, and it surpassed the others in recall and precision. It is the most suitable system to be deployed in remote forest regions due to its scalability, affordability, and efficiency. Active conservation and fire management practices are empowered by its real-time performance. Its AI and IoT synergy make it revolutionary in disaster prevention, with response time sped up and lowered monetary and environmental expenditure.