Integrating Multiple Attention Mechanism Fusion Based YOLO Logistics Sorting and Detection Model

In order to address the low efficiency of traditional manual logistics sorting and the low recognition accuracy and high computational complexity of traditional YOLO based visual recognition schemes in complex environments such as multi-scale detection and occlusion, this paper proposes an improved YOLO logistics automatic sorting algorithm that integrates multiple attention mechanisms. Based on the latest YOLOv11 network architecture, a small object detection layer is purposefully introduced to improve detection capability while controlling computational growth. Furthermore, a dual attention mechanism of channel and space is introduced to improve its performance in feature extraction and recognition accuracy. Combining the design of logistics sorting machinery equipment and software development, based on the logistics package dataset, the detection accuracy of different recognition schemes is compared. The experiment showed that compared with YOLOv8n, the proposed detection algorithm has better accuracy, recall mAP@0.5 Increase by 9.1%,4.3%, and 4.4% respectively.