Crucible defect recognition based on lightweight YOLOv8
LUO Binbin
PANG Qingwei
KANG Xiping
Abstract：To address the challenges of low detection reliability in robotic intelligent laboratory systems caused by sig-nificant geometric diversity of crucibles and imbalanced defect sample distribution,this paper proposes a crucible de-fect detection method based on the lightweight YOLOv8.Firstly,an image data augmentation strategy using Denoising Diffusion Probabilistic Model(DDPM)is introduced to balance and expand crucible image data according to the dis-tribution of data labels.Subsequently,to better balance the inference speed and detection accuracy of the network model,an improved YOLOv8 crucible defect detection algorithm is proposed,incorporating a non-parametric atten-tion mechanism and an adaptive weighting scheme.Experimental results demonstrate that the proposed method a-chieves a mean Average Precision(mAP)of 97.40%with an inference time of 1.20 ms per image.Although the in-ference speed is 0.2 ms slower than YOLOv11-n,the detection accuracy is optimal among current general object de-tection methods,especially lightweight ones.The proposed method can be effectively applied in industrial crucible defect detection scenarios.
Keywords：crucible defect recognition algorithmrobotic intelligent laboratory systemYOLOv8data augmenta-tionobject detectionweighted mechanisminference time
Publication Date：2025-03-31
Pages：7( 85-90,96 )