Infrared moving target detection system based on improved YOLOv8

According to the improved YOLOv8-based infrared moving target detection system provided by the invention, the double-input processing module controls the opening and closing of the optical flow processing module by using the detection confidence coefficient, the characteristics of the primary input and the secondary input are effectively fused together, the double-input processing module replaces the input of a YOLOv8 model, and the detection accuracy is improved. In this way, the main input and the auxiliary input can be efficiently fused together, the use of two YOLO network processing features is avoided, and only one YOLO network needs to be used after fusion; moreover, whether the auxiliary input is started or not is controlled by using the detection confidence coefficient, and when the detection confidence coefficient is high, the detection task can be well completed only by single input, so that the auxiliary input is not needed, the computing power can be saved to the greatest extent, and the real-tim