FIRE-DET:一种高效的火焰检测模型  

FIRE-DET:an efficient flame detection model

作　　者：陈浩霖 高尚兵 相林[1] 蔡创新 汪长春 CHEN Haolin;GAO Shangbing;XIANG Lin;CAI Chuangxin;WANG Changchun(Faculty of Computer and Software Engineering,Huaiyin Institute of Technology,Huaian 223001;Laboratory for Internet of Things and Mobile Internet Technology of Jiangsu Province,Huaiyin Institute of Technology,Huaian 223001)

机构地区：[1]淮阴工学院计算机与软件工程学院,淮安223001 [2]淮阴工学院江苏省物联网移动互联技术工程实验室,淮安223001

出　　处：《南京信息工程大学学报（自然科学版）》2023年第1期76-84,共9页Journal of Nanjing University of Information Science & Technology（Natural Science Edition）

基　　金：国家重点研发计划(2018YFB1004904);江苏省高校自然科学研究重大项目(18KJA520001);2021年淮阴工学院研究生科技创新计划项目(HGYK202122)。

摘　　要：模型的效率在计算机视觉中变得越来越重要.本文通过研究用于火焰检测的神经网络结构,提出了几个关键的优化方案,以提高模型效率和检测效果.第一,提出一种由多卷积组合结构构建的主干网络(FIRE-Net),它能高效地从多个尺度上提取丰富的火焰特征;第二,提出一种改进的加权双向特征金字塔网络(BiFPN-mini)以快速地实现多尺度特征融合;第三,提出一种新的注意力机制(FIRE-Attention),让检测器对火焰特征更敏感.基于上述优化,本文开发出了一种全新的火焰检测器FIRE-DET,它在硬件资源有限的条件下能够取得比现有基于深度学习的火焰检测方法更高的检测效率.FIRE-DET模型在自建数据集上进行训练后,最终对火焰检测的准确率和帧率分别达到97%和85 FPS.实验结果表明,与主流算法相比,本文火焰检测模型检测性能更优.本文为解决火焰探测问题提供了一个更通用的解决方案.In view of the increasing concern on model efficiency in computer vision, this paper proposed several optimization schemes to improve the flame detection models in model efficiency as well as the detection performance.A backbone network(FIRE-Net) was constructed from a multi-convolution combined structure, which can efficiently extract rich flame features from multiple scales.Then an improved weighted bidirectional feature pyramid network(BiFPN-mini) was used to quickly achieve multi-scale feature fusion.In addition, a new attention mechanism(FIRE-Attention) was proposed to make the detector more sensitive to flame characteristics.The above optimizations were combined to develop a new flame detector abbreviated as FIRE-DET,which was then trained on self-built dataset and tested on internet videos.The experimental results showed that the FIRE-DET outperformed mainstream algorithms by its flame recognition accuracy of 97% and frame rate of 85 FPS,thus provides a more common solution to solve the flame detection.

关 键 词：特征提取 特征融合 注意力机制 火焰检测 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]