Journal of Computer Applications ›› 2020, Vol. 40 ›› Issue (5): 1470-1475.DOI: 10.11772/j.issn.1001-9081.2019101737

• Virtual reality and multimedia computing • Previous Articles     Next Articles

Faster R-CNN based color-guided flame detection

HUANG Jie1, CHAOXIA Chenyu2, DONG Xiangyu1, GAO Yun3, ZHU Jun1, YANG Bo1, ZHANG Fei2, SHANG Weiwei2   

  1. 1.Maintenance Branch, State Grid Anhui Electric Power Company Limited, HefeiAnhui 230061, China
    2.School of Information Science and Technology, University of Science and Technology of China, HefeiAnhui 230000, China
    3.State Grid Anhui Electric Power Company Limited, HefeiAnhui 230022, China
  • Received:2019-10-14 Revised:2019-12-09 Online:2020-05-10 Published:2020-05-15
  • Contact: SHANG Weiwei, born in 1981, Ph. D., associate professor. His research interests include intelligent robot.
  • About author:DONG Xiangyu, born in 1981, M. S., engineer. His research interests include management of substation operation and maintenance.GAO Yun, born in 1977, M. S., engineer. Her research interests include management of substation operation and maintenance.ZHU Jun, born in 1987, engineer. His research interests include management of substation operation and maintenance.YANG Bo, born in 1988, engineer. His research interests include management of substation operation and maintenance.ZHANG Fei, born in 1991, Ph. D., engineer. His research interests include structural optimization design and parameter calibration of parallel robots.SHANG Weiwei, born in 1981, Ph. D., associate professor. His research interests include intelligent robot.
  • Supported by:

    This work is partially supported by the Scientific and Technical Project of State Grid Anhui Electric Power Company Limited (52120319000A).

基于Faster R-CNN的颜色导向火焰检测

黄杰1, 巢夏晨语2, 董翔宇1, 高云3, 朱俊1, 杨波1, 张飞2, 尚伟伟2   

  1. 1.国网安徽省电力有限公司 检修分公司, 合肥 230061
    2.中国科学技术大学 信息科学技术学院,合肥 230027
    3.国网安徽省电力有限公司,合肥 230022
  • 通讯作者: 尚伟伟(1981—)
  • 作者简介:黄杰(1988—),男,湖南益阳人,工程师,主要研究方向:智能检测; 巢夏晨语(1996—),男,安徽合肥人,硕士研究生,主要研究方向:深度学习、计算机视觉; 董翔宇(1981—),男,安徽蚌埠人,工程师,硕士,主要研究方向:变电站运维管理; 高云(1977—),女,河北石家庄人,工程师,硕士,主要研究方向:变电站运维管理; 朱俊(1987—),男,广西柳州人,工程师,主要研究方向:变电站运维管理; 杨波(1988—),男,陕西安康人,工程师,主要研究方向:变电站运维管理; 张飞(1991—),男,安徽合肥人,工程师,博士,主要研究方向:并联机器人结构优化设计与参数标定; 尚伟伟(1981—),男,江西南昌人,副教授,博士,主要研究方向:智能机器人。
  • 基金资助:

    国网安徽省电力有限公司2019年科技项目(52120319000A)。

Abstract:

Aiming at the problem of low detection rate of depth feature based object detection method Faster R-CNN (Faster Region-based Convolutional Neural Network) in flame detection tasks, a color-guided anchoring strategy was proposed. In this strategy, a flame color model was designed to limit the generation of anchors, which means the flame color was used to limit the generation locations of the anchors, thereby reducing the number of initial anchors and improving the computational efficiency. To further improve the computational efficiency of the network, the masked convolution was used to replace the original convolution layer in the region proposal network. Experiments were conducted on BoWFire and Corsician datasets to verify the detection performance of the proposed method. The experimental results show that the proposed method improves detection speed by 10.1% compared to the original Faster R-CNN, has the F-measure of flame detection of 0.87 on BoWFire, and has the accuracy reached 99.33% on Corsician.The proposed method can improve the efficiency of flame detection and can accurately detect flames in images.

Key words: fire detection, color model, Convolutional Neural Network (CNN), Faster Region-based Convolutional Neural Network (Faster R-CNN), anchor

摘要:

基于深度特征的目标检测方法Faster R-CNN在火焰检测任务上存在检测效率低的问题,因此提出了基于颜色引导的抛锚策略。该策略设计火焰颜色模型来限制锚的生成,即利用火焰颜色约束锚的生成区域,从而减少了初始锚的数量,提升了计算效率。为了进一步提高网络的计算效率,将区域生成网络中的卷积层替换成掩膜卷积。为了验证所提方法的检测效果,采用BoWFire和Corsician数据集进行验证。实验结果表明,该方法实际检测速度相较于原Faster R-CNN提高了10.1%,BoWFire上该方法的火焰检测F值为0.87,Corsician上该方法的准确度可达99.33%。所提方法可以提高火焰检测的效率,并能够准确检测图像中的火焰。

关键词: 火焰检测, 颜色模型, 卷积神经网络, Faster R-CNN,

CLC Number: