![Sample our Computer Science journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5928/TOC-Subject-Sample-2021-Computer-Science.jpg"})
ABSTRACT
Due to factors such as smoke, the non fire identification rate and fire identification accuracy in building fire identification are low, resulting in poor identification results. Therefore, a simplified identification method for building fire spread risk based on digital image processing technology is proposed. A fire spread model is constructed by considering both thermal radiation and thermal plumes, and based on this model, changes in smoke, temperature, and other factors during the occurrence of a fire are obtained. Using K-means to improve the dark channel defogging algorithm, the image after defogging is obtained, making the colour of the image more prominent and the feature information more abundant. Using the area threshold method to further extract the flame target contour in the defogged image, and determine whether the suspected flame area is a flame. Finally, four aspects of colour, sharp corner characteristics, area growth characteristics, and automatic estimation of propagation direction were identified to obtain simplified identification results for fire propagation risk. The experimental results show that the average non fire recognition rate of the proposed method reaches 95.8%, with good fire recognition accuracy and recognition effect, which verifies its feasibility.
Acknowledgments
The study was supported by ‘Construction and management of the experiment and training platform of production-education integration in higher vocational colleges based on “Internet +”(Grant No.19)’.
Disclosure statement
No potential conflict of interest was reported by the author(s).