For the characteristics of small,dense distribution,high diversity of defects and high precision and fast detection in the process of PCB(Printed Circuit Board)defect detection,a defect detection algorithm based on YO...For the characteristics of small,dense distribution,high diversity of defects and high precision and fast detection in the process of PCB(Printed Circuit Board)defect detection,a defect detection algorithm based on YOLOv8n was proposed in this study.Firstly,the original C2f module of YOLOv8n was improved into a C2FFaster-EMA module to reduce the number of parameters and floating-point operations(FLOPs).Additionally,the WIoUv3 loss function was introduced to mitigate the negative impact of low-quality defect images on model training.Consequently,a reduction in model size and an enhancement in detection precision were achieved.Finally,the ablation and comparative experiments were conducted on an augmented Deep PCB dataset,and the generalization experiments were performed on the PCB Defect-Augmented dataset.The results indicated that the proposed model reduces the number of parameters by 23.3%and FLOPs by 20%,P by 0.7%,mAP@0.5 by 0.3%,and mAP@0.5:0.95 by 3.9%,respectively,compared to the original YOLOv8n model.Furthermore,the comparative experiments demonstrated that the proposed model achieves higher accuracy and mAP compared to YOLOv5n and YOLOv5s.It was concluded that the proposed method satisfies the requirements for both accuracy and speed in PCB defect detection.展开更多
To advance the printing manufacturing industry towards intelligence and address the challenges faced by supervised learning,such as the high workload,cost,poor generalization,and labeling issues,an unsupervised and tr...To advance the printing manufacturing industry towards intelligence and address the challenges faced by supervised learning,such as the high workload,cost,poor generalization,and labeling issues,an unsupervised and transfer learning-based method for printing defect detection was proposed in this study.This method enabled defect detection in printed surface without the need for extensive labeled defect.The ResNet101-SSTU model was used in this study.On the public dataset of printing defect images,the ResNet101-SSTU model not only achieves comparable performance and speed to mainstream supervised learning detection models but also successfully addresses some of the detection challenges encountered in supervised learning.The proposed ResNet101-SSTU model effectively eliminates the need for extensive defect samples and labeled data in training,providing an efficient solution for quality inspection in the printing industry.展开更多
A magnet is an important component of a speaker,as it makes the coil move back forth,and it is commonly used in mobile information terminals.Defects may appear on the surface of the magnet while cutting it into smalle...A magnet is an important component of a speaker,as it makes the coil move back forth,and it is commonly used in mobile information terminals.Defects may appear on the surface of the magnet while cutting it into smaller slices,and hence,automatic detection of surface cutting defect detection becomes an important task for magnet production.In this work,an image-based detection system for magnet surface defect was constructed,a Fourier image reconstruction based on the magnet surface image processing method was proposed.The Fourier transform was used to get the spectrum image of the magnet image,and the defect was shown as a bright line in it.The Hough transform was used to detect the angle of the bright line,and this line was removed to eliminate the defect from the original gray image;then the inverse Fourier transform was applied to get the background gray image.The defect region was obtained by evaluating the gray-level differences between the original image and the background gray image.Further,the effects of several parameters in this method were studied and the optimized values were obtained.Experiment results show that the proposed method can detect surface cutting defects in a magnet automatically and efficiently.展开更多
To address the problems of large number of parameters and high complexity of calculation in the current steel surface defect detection model,a steel surface defect lightweight algorithm CGV-YOLO based on the improveme...To address the problems of large number of parameters and high complexity of calculation in the current steel surface defect detection model,a steel surface defect lightweight algorithm CGV-YOLO based on the improvement of YOLOv8 was proposed in this study.Firstly,in the process of optimizing the network architecture,the algorithm designed the FRC module and embeds it in the backbone network.Then,the GSConv convolution was employed to construct the Slim-neck network architecture,which further reduces computational load while maintaining model accuracy.Finally,the optimized CBAM replaced the C2f module in the YOLOv8 backbone network,reducing both model parameters and computational load.Based on the database of NEU-DET and BSD,the F1-Score of the CGV-YOLO algorithm is improved by 1.3%and 1.1%respectively compared with the baseline model.Based on the database of NEU-DET,the Params and computational complexity of the model are reduced by 30.6%and 35.3%respectively against the baseline.The results demonstrated that the proposed algorithm drastically reduces the number of parameters and computational cost with the maintenance of the accuracy of the model and realizes the lightweight effect.展开更多
文摘For the characteristics of small,dense distribution,high diversity of defects and high precision and fast detection in the process of PCB(Printed Circuit Board)defect detection,a defect detection algorithm based on YOLOv8n was proposed in this study.Firstly,the original C2f module of YOLOv8n was improved into a C2FFaster-EMA module to reduce the number of parameters and floating-point operations(FLOPs).Additionally,the WIoUv3 loss function was introduced to mitigate the negative impact of low-quality defect images on model training.Consequently,a reduction in model size and an enhancement in detection precision were achieved.Finally,the ablation and comparative experiments were conducted on an augmented Deep PCB dataset,and the generalization experiments were performed on the PCB Defect-Augmented dataset.The results indicated that the proposed model reduces the number of parameters by 23.3%and FLOPs by 20%,P by 0.7%,mAP@0.5 by 0.3%,and mAP@0.5:0.95 by 3.9%,respectively,compared to the original YOLOv8n model.Furthermore,the comparative experiments demonstrated that the proposed model achieves higher accuracy and mAP compared to YOLOv5n and YOLOv5s.It was concluded that the proposed method satisfies the requirements for both accuracy and speed in PCB defect detection.
文摘To advance the printing manufacturing industry towards intelligence and address the challenges faced by supervised learning,such as the high workload,cost,poor generalization,and labeling issues,an unsupervised and transfer learning-based method for printing defect detection was proposed in this study.This method enabled defect detection in printed surface without the need for extensive labeled defect.The ResNet101-SSTU model was used in this study.On the public dataset of printing defect images,the ResNet101-SSTU model not only achieves comparable performance and speed to mainstream supervised learning detection models but also successfully addresses some of the detection challenges encountered in supervised learning.The proposed ResNet101-SSTU model effectively eliminates the need for extensive defect samples and labeled data in training,providing an efficient solution for quality inspection in the printing industry.
基金Project (51575542) supported by the National Natural Science Foundation of ChinaProject (2016CX010) supported by the Innovation-Driven Project of CSU,ChinaProject (2015CB057202) supported by the National Basic Research Program of China
文摘A magnet is an important component of a speaker,as it makes the coil move back forth,and it is commonly used in mobile information terminals.Defects may appear on the surface of the magnet while cutting it into smaller slices,and hence,automatic detection of surface cutting defect detection becomes an important task for magnet production.In this work,an image-based detection system for magnet surface defect was constructed,a Fourier image reconstruction based on the magnet surface image processing method was proposed.The Fourier transform was used to get the spectrum image of the magnet image,and the defect was shown as a bright line in it.The Hough transform was used to detect the angle of the bright line,and this line was removed to eliminate the defect from the original gray image;then the inverse Fourier transform was applied to get the background gray image.The defect region was obtained by evaluating the gray-level differences between the original image and the background gray image.Further,the effects of several parameters in this method were studied and the optimized values were obtained.Experiment results show that the proposed method can detect surface cutting defects in a magnet automatically and efficiently.
文摘To address the problems of large number of parameters and high complexity of calculation in the current steel surface defect detection model,a steel surface defect lightweight algorithm CGV-YOLO based on the improvement of YOLOv8 was proposed in this study.Firstly,in the process of optimizing the network architecture,the algorithm designed the FRC module and embeds it in the backbone network.Then,the GSConv convolution was employed to construct the Slim-neck network architecture,which further reduces computational load while maintaining model accuracy.Finally,the optimized CBAM replaced the C2f module in the YOLOv8 backbone network,reducing both model parameters and computational load.Based on the database of NEU-DET and BSD,the F1-Score of the CGV-YOLO algorithm is improved by 1.3%and 1.1%respectively compared with the baseline model.Based on the database of NEU-DET,the Params and computational complexity of the model are reduced by 30.6%and 35.3%respectively against the baseline.The results demonstrated that the proposed algorithm drastically reduces the number of parameters and computational cost with the maintenance of the accuracy of the model and realizes the lightweight effect.
