A Flame Detection Algorithm Based on Improved greenYOLOv7

Flame recognition is of great significance in fire prevention. However, current algorithms for flame detection have some problems, such as missing and false detection, the accuracy cannot satisfy requirements In order to further above we propose a algorithm based on an improved YOLOv7 network. our algorithm, replace convolution MP-1 module with SimAM structure, which parameter-free attention mechanism. this way, problem can be improved. Furthermore, use ConvNeXt-based CNeB ELAN-W increasing complex environments. Finally, evaluate performance through large number test cases, data set used experiments was constructed by combining several publicly available sets various application scenarios. The experimental results indicate that compared original proposed achieve 7% increase aspect mAP_0.5 4.1% F1 score.