关键词: 消防, 神经网络, 嵌入式系统, 烟雾检测

Abstract: The traditional fire detection system used on portable devices consumes a lot of memory, is vulnerable to the environment, and the accuracy needs to be improved. We propose a fire detection system with a small memory footprint and excellent performance. Using the FireNet neural network we designed to embed in portable hardware products, we tested it on a public data set. Compared with other traditional methods, the test results have significantly improved the accuracy and speed.

Key words: fire Protection, neural networks, embedded systems, smoke detection