Design and Research on Automatic Fire Alarm System for Subway Based on Visual Flame Detection Algorith
To improve the accuracy of subway fire detection,an improved Yolov 3 visual flame detection algorithm is proposed,and an automatic subway fire alarm system is designed based on this algorithm.The visual flame detection algorithm is based on Yolov 3 as the structural foundation,and the backbone network is designed as multiple sets of multi-scale convolution patterns to ensure the integrity of the extracted information.Hollow convolution calculation is added to simplify the calculation parameters and feature extrac-tion range;According to the idea of pyramid networks,the information fusion part of the strengthened neural network is set to a bidi-rectional fusion structure to ensure the sensitivity of the algorithm to small flames.The subway fire automatic alarm system uses FPGA chips to build the main control module,and wireless communication is carried out according to the ZigBe protocol,which has good op-erating speed and information transmission speed.Finally,experiments have shown that the visual flame detection algorithm can de-tect small fires with minimal flames,and the fire detection rate is high and the detection accuracy is good;The subway fire automatic alarm system based on visual flame detection algorithm can operate normally and achieve monitoring and alarm of subway fires,basi-cally meeting the design requirements.
inception networkpyramid networkYolov 3wireless communication
万方数据
维普
