Substation equipment detection method and system based on bimodal data fusion

The invention discloses a transformer substation equipment detection method and system based on bimodal data fusion, and the method comprises the following steps: obtaining an infrared modal image and a visible light modal image of to-be-detected transformer substation equipment from a transformer substation, and carrying out the preprocessing of the infrared modal image and the visible light modal image; based on the double-flow-YOLOv5 neural network, a multi-modal detection model is constructed; inputting the preprocessed infrared modal image and the preprocessed visible light modal image into a multi-modal detection model to obtain an infrared modal feature map and a visible light modal feature map respectively; carrying out self-attention fusion, and adding a series of feature maps to obtain a fused feature map; obtaining target information based on the fused feature map; and performing target positioning on the target information to obtain a target position, and completing detection of the substation equ