Study on flame detection based on Yolo neural network and machine vision

Journals

Online: 29 August 2022; Volume 1, Issue 1, No.2 PDF DOWNLOAD

Title：
Study on flame detection based on Yolo neural network and machine vision
Author：
Si Chen , Tao-chuan Zhang , Xiu-quan Liu , Ming Yi
Author Affiliation：
Intelligent Manufacturing College of Foshan Polytechnic, Foshan, China

Abstract

To improve reliability of the fire pre-warning system, study on flame detection was carried out in combination with YOLO artificial neural network and machine vision detection technology. Firstly, 200 flame pictures were acquired and 1,000 flame imaged were searched online to serve as the training set, and the flame area of each image was labelled artificially, and then YOLOv5s neural network model with small scale and rapid detection was built, and the system was trained by means of supervised learning, and at last, the trained system was used to make real-time detection of the flame. The experiment results showed that, the system could make real-time monitoring of flames, with relatively high recognition accuracy, rapid detection speed and high recognition flexibility.

References

[1] LIU Wei-peng, CAO Yu-dong. Forest Fire Image Recognition Algorithm Based on In-depth Learning and its Implementation [J]. Changjiang Information of Communications, 2022, 35(1):29-31.

[2] ZHAO Yuan-yuan, ZHU Jun, XIE Ya-kun, et al. Real-time Detection Algorithm of Video Image Flame Based on YOLOv3 [J]. Journal of Wuhan University (Information Science Edition), 2021, 46(3):326-334.

[3] GU Fang-ying. Study and Application of Fuzzy Neural Network in Detection of Shopping Mall Fire [D]. Huainan: Anhui University of Science and Technology, 2019.

[4] LIU Tong-jun, NIAN Fu-dong, LV Gang, et al. Image Flame Detection Method by Super-pixel Positioning Guidance based on FasterR-CNN [J]. Journal of Hefei Normal University, 2020, 38(6):29-32.

[5] MA Lin-lin, MA Jian-xin, HAN Jia-fang, et al. Study on Object Detection Algorithm Based on YOLOv5s [J]. Computer Knowledge and Technology, 2021, 17(23):100-103.

[6] YU Yang, ZHOU Luo-yu, XIONG Li-ya, et al. Infusion Bottle Object Detection based on Improved YOLOv5s [J]. Electronics World, 2022(1):182-183+197.

[7] LI Zhi-gang, ZHANG Na. A Lightweight YOLOv5 Traffic Sign Recognition Method [J/OL]. Telecommunications Technology: 6 1-7(2021-12-31)[2022-01-03].http://kns. cnki.net/kcms/detail/51.1267.tn.20211230.1940.014.html.

[8] DENG Tian-min, TAN Si-qi, PU Long-zhong. Study on Recognition Method of Traffic Light Based on Improved YOLOv5 [J/OL]. Computer Engineering: 1-13(2021-12-20)[2021-12-24].https://kns.cnki.net/kcms/detail/detail.aspx?dbcode=-CAPJ&dbname=CAPJLAST&filename=J SJC20211217001&uni-platform=NZKPT&v=g9BjGJf5ZLWeY9srs5mjc8fg0aR8Anl-R0NG5WbFaOCin9XALUJxVteFbas0Z 6xJE.

[9] YANG Xiao-ling, CAI Ya-wen. Pedestrian Detection System Based on yolov5s and its Implementation [J]. Computer and Information Technology, 2022, 30(1):28-30.

[10] LI Yi-ming, WANG Xiao. Surface Defect Detection of Rolled Steel based on YOLOv5s Model [J]. Manufacturing Automation, 2021, 43(11):117-119.

[11] YANG Yong-bo, LI Dong. Wearing Detection Algorithm of Lightweight Safety Helmet based on Improved YOLOv5 [J/OL]. Computer Engineering and Application: 1-8(2021-01-19)[2021-11-01]. http://kns.cnki.net/kcms/detail/11.2127.TP.20220118.1827.006.html.

[12] SHEN Ke, JI Liang, ZHANG Yuan-hao, et al. Coal Gangue Object Detection based on Improved YOLOv5s Model [J]. Journal of Mine Automation, 2021,47(11):107-111+118.