Reagent Container Detection Based on PC-YOLO Algorithm

In order to solve the problem of low detection accuracy of small targets caused by light change, background interference and occlusion, an improved algorithm based on YOLOv11, PC-YOLO, was proposed in this paper. The algorithm combines the C3K2 module and the PPA module of YOLOv11 to construct the C3K2_PPA module, effectively retaining the key features of small targets and reducing false detections. It introduces the CMUNeXt Block to replace traditional convolution operations, enhancing global information extraction through depthwise separable convolutions and residual connections. Additionally, the MPDIoU method is employed to optimize the precision of bounding box regression, accelerating the convergence of training. Experimental results demonstrate that, compared to the original YOLOv11, PC-YOLO improves mAP50 from 92.2% to 94.2% and mAP50-95 from 81.2% to 85.0%, achieving superior detection performance in complex environments.