Filter-PCA-Based Process Monitoring and Defect Identification During Climbing Helium Arc Welding Process Using DE-SVM

Hong, Yongmiao; Yang, Mingxuan; Chang, Baohua; Du, Dong

doi:10.1109/tie.2022.3201304

Cited by 18 publications

(2 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For handcrafted methods, one can emphasize the approach presented in [17] and [3]. Different image processing stages are applied to the image before extracting the features for the classifiers.…”

Section: Related Workmentioning

confidence: 99%

LoHi-WELD: A Novel Industrial Dataset for Weld Defect Detection and Classification, a Deep Learning Study, and Future Perspectives

Biasuz Block,

Dutra da Silva,

Eugnio Lazzaretti

et al. 2024

IEEE Access

View full text Add to dashboard Cite

The automated inspection of weld beads is of great importance for many industrial processes. Failures may cause a loss of mechanical resistance of the weld bead and compromise the manufactured part. Several methods have been proposed in the literature to address this problem, and recently, methods based on deep learning have gained prominence in terms of performance and applicability. However, such methods require vast and reliable datasets for different real defects, which have yet to be available in recent literature. Hence, this paper presents LoHi-WELD, an original and public database to address the problem of weld defect detection and classification of four common types of defects -pores, deposits, discontinuities, and stains -with 3,022 real weld bead images manually annotated for visual inspection, composed by low and high-resolution images, acquired from a Metal Active Gas robotic welding industrial process. We also explore variations of a baseline deep architecture for the proposed dataset based on a YOLOv7 network and discuss several case analyses. We show that a lightweight architecture, ideal for industrial edge devices, can achieve up to 0.69 of mean average precision (mAP) considering a fine-grained defect classification and 0.77 mAP for a coarse classification. Open challenges are also presented, promoting future research and enabling robust solutions for industrial scenarios. The proposed dataset, architecture, and trained models are publicly available on GitHub a .

show abstract

Section: Related Workmentioning

confidence: 99%

LoHi-WELD: A Novel Industrial Dataset for Weld Defect Detection and Classification, a Deep Learning Study, and Future Perspectives

Biasuz Block,

Dutra da Silva,

Eugnio Lazzaretti

et al. 2024

IEEE Access

View full text Add to dashboard Cite

show abstract

“…At the same time, it is also necessary to ensure that the particle size of the ceramic powder is not greater than 0.03mm and evenly distributed; However, if the particles are too dense, it may lead to a large gap between the green body and the ferrite line, resulting in poor product quality, waste of raw materials, or loss of raw materials during the production process, and even burning loss. In addition, ceramic defects are also an important factor affecting product quality [7][8].…”

Section: Porcelain Product Defect Categorymentioning

confidence: 99%

Dehua porcelain product defect recognition based on image processing

Gao

2023

Third International Conference on Advanced Algorithms and Signal Image Processing (AASIP 2023)

View full text Add to dashboard Cite

In recent years, with the continuous development of the ceramic industry, the problem of Product defect has become increasingly prominent. Therefore, this article conducted recognition research on Dehua porcelain through image processing technology. Firstly, the background and research significance of this article were elaborated; Secondly, a detailed analysis was conducted on the detection technology and application status of powder and particle size in ceramic processing; Next, the principle and implementation process of defect recognition based on machine vision were elaborated, and experimental results were compared and error corrected. At the same time, improved algorithms were proposed to improve identification accuracy; Finally, this paper tests the performance of the porcelain Product defect recognition model, and the test results show that the resource occupancy rate of this model is between 2% and 4%, and the defect recognition rate is relatively high, which can meet the demand. Compared with the traditional visual inspection, the defect recognition method based on image processing can realize the automatic detection process. It can improve production efficiency and reduce the impact of human errors and subjective judgments. This method can identify the defects of Dehua porcelain products in real time and reduce the time and labor cost of traditional detection methods.

show abstract

Assessing thermal properties of multipass weld beads using active thermography: microstructural variations and anisotropy analysis

Sesana,

Santoro,

Curà

et al. 2023

Int J Adv Manuf Technol

View full text Add to dashboard Cite

In the present paper, the results of a research focused on microstructure investigation of steel welded joints by means of non destructive thermographic technique are presented. In particular, two rough joints obtained by means of irregular manual welding are analyzed. Active lock in thermography phase analysis is applied, basing on the evidence that thermal diffusivity is related to the thermal diffusion length, using the phase contrast trend as a function of the distance from the laser spot. Diffusivity is expected to be dependent on the local microstructure and vary through different directions. By the variation in diffusivity, it is possible to investigate the level of anisotropy.The results are compared with micrographic analysis. The proposed method provides an active lock-in infrared thermography analysis by means of a thermal camera and an exciting laser head to evaluate thermal diffusivity.

show abstract

Filter-PCA-Based Process Monitoring and Defect Identification During Climbing Helium Arc Welding Process Using DE-SVM

Cited by 18 publications

References 26 publications

LoHi-WELD: A Novel Industrial Dataset for Weld Defect Detection and Classification, a Deep Learning Study, and Future Perspectives

LoHi-WELD: A Novel Industrial Dataset for Weld Defect Detection and Classification, a Deep Learning Study, and Future Perspectives

Dehua porcelain product defect recognition based on image processing

Assessing thermal properties of multipass weld beads using active thermography: microstructural variations and anisotropy analysis

Contact Info

Product

Resources

About