Surface Defect Detection using Hierarchical Features

Ling, Xiao; Huang, Tao; Wu, Bo; Hu, Youmin; Zhou, Jiehan

doi:10.1109/coase.2019.8843235

Cited by 5 publications

(2 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Feng [36] proposed an improved encoding-decoding network based on feature image fusion to detect cracks in hydroelectric dam images acquired by an unmanned aerial vehicle (UAV). Xiao [37] introduced a hierarchical-feature-based convolution neural network (H-CNN) model to detect oil leaks in freight trains. Due to the high level of standardization in industrial processes, instances of labeled damage patterns are seldom available.…”

Section: Introductionmentioning

confidence: 99%

Anomaly Detection in the Production Process of Stamping Progressive Dies Using the Shape- and Size-Adaptive Descriptors

Ma,

Meng

2023

Sensors

View full text Add to dashboard Cite

In the production process of progressive die stamping, anomaly detection is essential for ensuring the safety of expensive dies and the continuous stability of the production process. Early monitoring processes involve manually inspecting the quality of post-production products to infer whether there are anomalies in the production process, or using some sensors to monitor some state signals during the production process. However, the former is an extremely tedious and time-consuming task, and the latter cannot provide warnings before anomalies occur. Both methods can only detect anomalies after they have occurred, which usually means that damage to the die has already been caused. In this paper, we propose a machine-vision-based method for real-time anomaly detection in the production of progressive die stamping. This method can detect anomalies before they cause actual damage to the mold, thereby stopping the machine and protecting the mold and machine. In the proposed method, a whole continuous motion scene cycle is decomposed into a standard background template library, and the potential anomaly regions in the image to be detected are determined according to the difference from the background template library. Finally, the shape- and size-adaptive descriptors of these regions and corresponding reference regions are extracted and compared to determine the actual anomaly regions. The experimental results indicate that this method can achieve reasonable accuracy in the detection of anomalies in the production process of stamping progressive dies. The experimental results demonstrate that this method not only achieves satisfactory accuracy in anomaly detection during the production of progressive die stamping, but also attains competitive performance levels when compared with methods based on deep learning. Furthermore, it requires simpler preliminary preparations and does not necessitate the adoption of the deep learning paradigm.

show abstract

Section: Introductionmentioning

confidence: 99%

Anomaly Detection in the Production Process of Stamping Progressive Dies Using the Shape- and Size-Adaptive Descriptors

Ma,

Meng

2023

Sensors

View full text Add to dashboard Cite

show abstract

“…Recently, with the development of deep learning, several studies to detect defects using a convolutional neural network (CNN) have been proposed [12][13][14][15][16][17][18][19][20][21][22][23]. Xiao et al [12] proposed a hierarchical feature-based CNN (H-CNN) structure that generates regions of interest (ROIs) using region-based CNN (R-CNN) and detects the defect using a fully CNN (F-CNN). Singh et al [14] compared the classification performance for STAIN defects on metal surfaces using a CNN such as ResNet [24] and YOLO [25].…”

Section: Introductionmentioning

confidence: 99%

STAIN Defect Classification by Gabor Filter and Dual-Stream Convolutional Neural Network

Kim²,

Park³

2022

Preprint

View full text Add to dashboard Cite

A STAIN defect is difficult to detect with the naked eye because of its characteristic of having a very minimal difference in brightness with the local area of the surface. Usually, background extraction–based and Gabor filter–based texture feature–based methods have been proposed to detect STAIN defects. Recently, with the development of deep learning, the convolutional neural network (CNN)–based detection method has been proposed. Gabor filter images have an advantage in image texture analysis and can be used as an input to the CNNs. However, because the original image and the Gabor filter image have different domains, it is necessary to design a CNN that can extract features advantageous for STAIN defect classification. In this paper, a CNN for STAIN defect classification using the original image and Gabor filter image was proposed. A Gary-Level Co-occurrence Matrix (GLCM) matrix from the Gabor filter image was constructed, and then the Inverse Difference Moment (IDM) value, which is a homogeneity local feature, was extracted. Gabor filter images suitable for STAIN defect classification were selected using IDM values, and a STAIN defect detection system that uses the selected Gabor filter images and the inspection image as an input to the CNN was constructed. As a result of the experiment based on the magnetic tile data set and the compact camera module (CCM) dataset, it was confirmed that the proposed method has an improved performance based on the precision, recall, and F1-score in comparison to the single-stream CNN-based method.

show abstract

Surface Defect Detection Using Image Pyramid

Ling

2020

IEEE Sensors J.

View full text Add to dashboard Cite

Surface Defect Detection using Hierarchical Features

Cited by 5 publications

References 31 publications

Anomaly Detection in the Production Process of Stamping Progressive Dies Using the Shape- and Size-Adaptive Descriptors

Anomaly Detection in the Production Process of Stamping Progressive Dies Using the Shape- and Size-Adaptive Descriptors

STAIN Defect Classification by Gabor Filter and Dual-Stream Convolutional Neural Network

Surface Defect Detection Using Image Pyramid

Contact Info

Product

Resources

About