Research on multi-camera data fusion for improving fire detection accuracy DOI Creative Commons
Wen Wang, Xiaochun Chen, Meng Zhou

и другие.

Applied Mathematics and Nonlinear Sciences, Год журнала: 2024, Номер 9(1)

Опубликована: Янв. 1, 2024

Abstract With the rapid urbanization in China, use of various electrical equipment and a large number flammable materials has led to an increasing trend frequency fires from year year. In this paper, we start with data fusion collect fire open fragments so as establish detection dataset. A monitoring terminal that utilizes multi-feature is created using algorithm convolutional neural network improve main structure YOLOv5 model. The effect improved model compared other models when combined. it found better training time steady state than three groups models, its mAP value by 22.1%, 13.6% 10.13% respectively. average accuracy for flames smoke generated different also higher models. At same time, stronger classification checking abilities, more accurate recognizing whether occurring image. improving model, effectively applied work, realizing dynamic analysis real-time flame providing effective monitoring.

Язык: Английский

Research on multi-camera data fusion for improving fire detection accuracy DOI Creative Commons
Wen Wang, Xiaochun Chen, Meng Zhou

и другие.

Applied Mathematics and Nonlinear Sciences, Год журнала: 2024, Номер 9(1)

Опубликована: Янв. 1, 2024

Abstract With the rapid urbanization in China, use of various electrical equipment and a large number flammable materials has led to an increasing trend frequency fires from year year. In this paper, we start with data fusion collect fire open fragments so as establish detection dataset. A monitoring terminal that utilizes multi-feature is created using algorithm convolutional neural network improve main structure YOLOv5 model. The effect improved model compared other models when combined. it found better training time steady state than three groups models, its mAP value by 22.1%, 13.6% 10.13% respectively. average accuracy for flames smoke generated different also higher models. At same time, stronger classification checking abilities, more accurate recognizing whether occurring image. improving model, effectively applied work, realizing dynamic analysis real-time flame providing effective monitoring.

Язык: Английский

Процитировано

0