Application of CARCAW-YOLOv11 Algorithm in Royal Jelly Impurity Detection

Royal jelly is a highly nutritious and health-beneficial bee product. However, its production process often involves impurities such as beeswax and bee pupae, which compromises its quality and value. Traditional methods for detecting impurities in royal jelly primarily rely on manual experience, leading to inefficiency and subjectivity, which fails to satisfy the demand for rapid and accurate detection. To address this, this paper proposed the CARCAW-YOLOv11 (Content-Aware Reinforced Coordinate Attention Wise-YOLOv11) algorithm for the rapid detection of royal jelly impurities. Based on the YOLOv11 architecture, the proposed method incorporated the CARAFE feature enhancement module, the Coordinate Attention (CA) mechanism, and an optimized Wise-IoU loss function. Experimental results indicated that the CARAFE module effectively reduced information loss during upsampling, providing higher-quality features for subsequent detection. The CA mechanism enhanced the model’s detection precision and recall, and the optimized Wise-IoU loss function improved the detection capability for small objects and reduced computational complexity. The CARCAW-YOLOv11 model achieved a precision of 91.7%, a recall of 80.5%, an mAP@0.5 of 87.3%, and an F1-score of 0.79, with a detection speed of 140.2 FPS. Compared with models such as YOLOv8n, YOLOv9s, Faster R-CNN, and DETR, CARCAW-YOLOv11 model demonstrated a significant improvement in detection accuracy for tiny targets and robustness in the royal jelly impurity detection. This study provided a novel, fast, and accurate approach for the quality control of royal jelly.