I. Introduction
The fusion of infrared and visible images is a hot topic in the field of dual-band object detection [1]. Visible light cameras can provide rich visual details in well-lit environments, capturing abundant color, texture, and detail information on the surface of objects. In contrast, infrared cameras are highly sensitive to temperature variations, making them suitable for special conditions such as nighttime and occlusions. They can detect object contours based on thermal radiation, unaffected by ambient light, which makes them significant for object detection. Even under low light or other harsh conditions like occlusion, fog, or smoke, they can still achieve good detection accuracy for vehicles, pedestrians, and other objects. In contrast, visible light images are more susceptible to the influence of ambient lighting. To address the problem of all-day operations, visible light images and infrared images are often fused or switched to address lighting issues [2], [3].