基于深度卷积神经网络的火灾预警算法研究

Research on Fire Warning Algorithm Based on Deep Convolutional Neural Network

烟雾检测在火灾预防中起着非常重要的作用,针对传统方法中检测精度低、开阔空间难以检测的问题,文章提出了基于深度卷积神经网络的火灾烟雾检测方法。近年来,卷积神经网络在各种目标检测中发挥着越来越重要的作用,各类应用于目标检测的神经网络框架也被提出并应用到实际生活当中,文章通过替换特征提取器(例如Inception Net和残差网络)和参数优化来改进现有的目标检测框架Faster R-CNN,Single Shot Multi Box Detector(SSD),Region-based Fully Convolutional Networks(R-FCN)并将其应用在火灾烟雾检测上。在烟雾检测数据上,m AP最高达到了56.04%,与现有的烟雾检测方法相比,文章方法在检测精度上和速度上都取得了较好的结果。

The smoke detection plays a very important role in prevention of fire. But the accuracy of smoke detection is low and difficult to detect in open space by traditional methods. In this paper, we introduce an improved object detection method based on deep convolution neural network（CNN） to address this issue. In recent years, various neural network frameworks for object detection have also been proposed and applied to real life.This paper improved the existing object detection framework such as Faster R-CNN, Single Shot Multi Box Detector（SSD）, Region-based Fully Convolutional Networks（R-FCN） by substituting the feature extractor（such as Inception Net and Res Net） and parameter optimization.The experiments result demonstrate the m AP reached 56.04% on the smoke detection dataset. Compared with the existing smoke detection methods, the method of the present invention has achieved good results both in smoke detection accuracy and speed.