/FireDetectionYOLOv2

Real-time Fire Detection for CCTV surveillance system using Deep Learning YOLOv2

Primary LanguageMATLAB

View Fire Detection for CCTV surveillance system using YOLOv2 on File Exchange

Fire Detection for CCTV surveillance system using Deep Learning


MATLAB program for CCTV surveillance system using Deep Learning

  • Needs from the construction industry
  • Applying deep learning to Video streams from CCTV
  • YOLOv2 deep learning model implemented to detect fire from video stream

Entire Workflow

  1. Access data : Acess 1000+ images with imagedatastore
  2. Labeling data : Automatic fire labeling class for image labeler defined using image processing apps, e.g. color thresholder, image segmenter
  3. Training : YOLOv2 training trainNetwork function
  4. Deployment : Inference speed acceleration by generating C/C++ CUDA mex file for real-time prediction

Dataset Used

  • Cazzolato, Mirela T., et al. "FiSmo: A Compilation of Datasets from Emergency Situations for Fire and Smoke Analysis." Proceedings of the satellite events (2017).
    Copyright 2019 The MathWorks, Inc.

Cite As
Wanbin Song (2019). Fire Detection for CCTV surveillance system using YOLOv2 (https://www.github.com/wanbin-song/FireDetectionYOLOv2), GitHub. Retrieved November 26, 2019.