YOLO4IIR: Infrared Image Recognition Model for Power Equipment Detection
Junhua Yin, Zhongfu Li, Feichi Zhong, Wentao Zhang, Qianyue Wang, Ganqguan Si · 2023
Object detection techniques have experienced rapid development. However, challenges such as low recognition accuracy, high false-negative rates and slow detection speed still exist in the task of infrared detection for power equipment. To address these issues, this paper proposes a power equipment infrared image recognition model (YOLO4IIR) based on attention mechanisms and convolutional neural networks. By improving upon the YOLOv5 architecture and introducing the EMO framework, CoTAttention, and CARAFE operator, the model's detection capability is enhanced while maintaining inference speed. The model is trained and tested on a custom infrared image dataset from a substation. The results demonstrate that the YOLO4IIR model outperforms the YOLOv5 model in terms of accuracy and other performance metrics, with little to no sacrifice in inference speed. This research provides an efficient and accurate infrared image recognition method for power equipment detection and fault diagnosis.