神经网络方法在变电站消防管理的应用研究

为了提升变电站消防管理科学化水平，实现视频图片火焰特征提取及识别定位。接入视频监控码流数据，构建基于卷积神经网络的火焰识别模型，进行实时识别火焰特征并预警预报。实验表明，该方法能够自动提取火焰特征，有效提高复杂背景下的火焰识别的准确率，具有良好的鲁棒性和泛化能力，在变电站消防管理中有较大应用前景。

Study on the application of neural network on transformer substation fire management

In order toimprove the fire management of transformer substation, the video and imageflame feature extraction and identification are achieved. The stream data ofvideo surveillance is accessed. The flame identification model based on convolutionneural network is built. The flame features are real time identified, and earlywarning and forecast are performed. Experiments showed that, the method canextract the flame feature automatically, and efficiently improve the accuracyof flame identification with complex background, with good robustness andgeneralization ability, and have great application prospect in the firemanagement of transformer substation.