DEVELOPMENT OF AN AERIAL FIRE IDENTIFICATION SYSTEM BASED ON VISUAL ARTIFICIAL INTELLIGENCE

• Main Author: Afifa, Hasna
• Format: Final Project
• Language: Indonesia
• Subjects: Teknik (Rekayasa, enjinering dan kegiatan berkaitan)
• Online Access: https://digilib.itb.ac.id/gdl/view/69266
• Institution: Institut Teknologi Bandung

DKI Jakarta is the most populous province in Indonesia. Based on statistics, in the capital city of this country there have been around 6,429 fires throughout 2020. This dense area makes the distance between houses very close which has the potential to accelerate the spread of fire from one house to another. To reduce losses due to fire, it is necessary to extinguish and rescue immediately. However, the fire trucks were unable to reach the fire site due to the narrow road access. In this case, the use of autonomous drones in collaboration with fire detection based on visual artificial intelligence can help fire disaster management. In the long term, it is hoped that drones can fly by themselves to the point of fire and then release fire-fighting bombs automatically. This undergraduate assignment explores the idea of identifying fire using a computer vision approach. Because the speed of identification is very important in disaster management cases, the model in this study uses YOLOv4 algorithm trained in Google Colaboratory, each model takes 8-10 hours to train. In this study, 8 identification models were built with each dataset of day, night, day and night, thermal, day filter, night filter, day and night filter, and thermal filter. Each dataset contains 300-450 data in the form of images from the top view in a state of fire and not fire, also at the daylight and night. The results show that the model with the day and night dataset is the most stable model among the other seven models, because this model has the highest average F1- score for each test.