An Improved Lightweight Common Object Detection Algorithm based on YOLO11

Aiming at the integration deployment requirements of mobile robots, this paper proposes a target recognition algorithm based on improved YOLO11. Firstly, the MSCV module is fused with the C3k2 module and introduced into the YOLO11 model to realize multi-scale feature extraction, so as to improve the object detection performance of the model. Secondly, the LDCM module is adopted to realize model lightweighting with negligible loss of detection accuracy. Finally, the PIoU v2 loss function is utilized to strengthen the target localization capability of the model. Compared with the original YOLO11 baseline model, the precision P, mAP@50 and FPS of the improved model are increased by 2.9%, 1.81% and 19.88%, respectively; the number of parameters and model complexity are reduced by 22.75% and 7.98%. The results demonstrate that the proposed improved model effectively boosts object detection performance while achieving lightweight design.