The Application of Artificial Intelligence-based Fireworks Recognition Technology in Fire Detection

• Published in: 2022 IEEE Asia-Pacific Conference on Image Processing, Electronics and Computers (IPEC) pp. 1371 - 1374
• Main Author: Yao, Xunxun
• Format: Conference Proceeding
• Language: English
• Published: IEEE 14.04.2022
• Subjects: Application Research; Artificial Intelligence; Computers; Conferences; Economics; Fire Detection; Firework Recognition; Image recognition; Neural networks; Urban areas; Wireless communication
• DOI: 10.1109/IPEC54454.2022.9777461

Fire has the characteristics of suddenness and randomness. It not only has a high probability of accidents, but also causes great damage to the environment. How to quickly and effectively implement fire-related alarm measures and significantly reduce fire-related losses has become an important part of fire prevention and research. Firework detection is one of the obvious physical characteristics that are easy to be inspected in the early stage of a major fire. Therefore, firework detection is an important subject with theoretical and practical value. This article mainly focuses on the actual application of artificial intelligence-based wireless pyrotechnic identification technology in fire detection. First, the literature research method is used to explain the pyrotechnic identification process and the fire process of the fire, and then the desired detection technology is introduced. Finally, the firework recognition experiment is mainly to compare the accuracy of artificial intelligence firework recognition technology. The experimental results show that the effect of pure application for flame recognition is very unsatisfactory, especially under long-distance conditions, the accuracy rate is only about 60%, the recognition effect of BP neural network is obviously better than that of Bayesian recognition.