The Fire Detector Model is an advanced system developed as part of the Quality Control and Computer Vision course. This project integrates cutting-edge computer vision techniques with temperature monitoring to achieve real-time fire detection. The model is trained using a ResNet-50 convolutional neural network, a deep learning architecture known for its accuracy in image recognition tasks. By leveraging a combination of visual data and input from a temperature sensor, the system ensures high reliability in identifying fire incidents.
The hardware implementation utilizes an Arduino microcontroller, which processes sensor data and interacts with the detection algorithm. When a fire is detected, the system immediately triggers a buzzer alarm, providing an audible alert to ensure timely intervention. This innovative project demonstrates the practical application of machine learning and embedded systems in safety-critical scenarios, offering a robust and efficient solution for fire monitoring and prevention.
Datasets used
- https://www.kaggle.com/datasets/atulyakumar98/test-dataset
- https://www.kaggle.com/datasets/metinmekiabullrahman/fire-detection
- https://www.kaggle.com/datasets/elmadafri/the-wildfire-dataset
COLE, Rob Mark. Fire detection from images. Disponível em: https://github.com/robmarkcole/fire-detection-from-images. Acesso em: 13 dez. 2024.
FIRE detection model training approach. Reddit: Computer Vision. Disponível em: https://www.reddit.com/r/computervision/comments/1g41cok/fire_detection_model_trai ning_approach/. Acesso em: 12 dez. 2024.
INSTITUTO NACIONAL DE ESTATÍSTICA (INE). Indicador: População residente com 15 e mais anos por nível de escolaridade completo. Disponível em: https://www.ine.pt/xportal/xmain?xpid=INE&xpgid=ine_indicadores&indOcorrCod=00 13509&contexto=bd&selTab=tab2. Acesso em: 28 dez. 2024.
Ala’ Khalifeh, AbdelHamid Nassar, Mohammad M. AlAjlouni, Anas AlNabelsi, Zaid Alrawashdeh, Bashar Hejazi, Radi Alwardat, Jose Lima . A Machine Learning-Based Early Forest Fire Detection System Utilizing Vision and Sensors’ Fusion Technologies. Disponível em: https://bibliotecadigital.ipb.pt/bitstream/10198/27446/1/Learning.pdf. Acesso em: 13 dez. 2024.
WASIKE, Bravin. Building a deep learning model with Keras and ResNet-50. Medium. Disponível em: https://medium.com/@bravinwasike18/building-a-deep-learning-model-with-keras-andresnet-50-9dd6f4eb3351. Acesso em: 15 dez. 2024.