基于LSTM网络的船舶机舱火灾多特征融合探测

针对船舶机舱火灾高效准确探测的需求，建立基于LSTM-ID3 判决的船舶火灾探测方法。首先确定采集船舶火灾特征的三类传感器，然后完成 LSTM 神经网络模型的构建、参数的优化，将 LSTM 神经网络输出的明火、阴燃火、无火的概率值与烟雾持续时间作为决策树的输入量，输出火灾探测结果。利用国家标准火典型数据进行训练，并开展相关试验，对船舶机舱火灾进行探测。试验结果表明，与其他算法进行对比，探测准确率达到97%以上，该方案能对机舱火灾做出有效探测，为船舶安全提供科学依据。

Multi feature fusion prediction of marine engineroom fire based on LSTM network

In view of the demand for efficient and accurate prediction of the fire in the engine room of the ship, this paper establishes a ship fire prediction method based on the LSTM neural network and decision tree decision. Firstly, it determines three kinds of sensors to collect the fire characteristics of the ship. Secondly, it constructs the LSTM neural network model and optimizes the parameters. The probability of open fire, smoldering fire, no fire and smoke duration output by STM neural network are the inputs of decision tree, and the fire prediction results are output. The typical data of national standard fire are used for training, and experiments are conducted to predict the fire in the engine room of the ship. The experimental results show that the prediction accuracy is over 97% compared with other algorithms. The scheme can effectively predict engine room fire and provide scientific basis for ship safety.