Automated vision system for magnetic particle inspection of crankshafts using convolutional neural networks

Tout, Karim; Meguenani, Anis; Urban, Jean-Philippe; Cudel, Christophe

doi:10.1007/s00170-020-06467-4

Cited by 25 publications

(20 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The method of combining statistical features with machine learning is mainly to extract statistical features from the defect surface, and then use machine learning algorithms to learn these features to achieve surface defect detection. In recent years, with the success of the deep learning model represented by the convolutional neural network (CNN) in many computer vision fields, the defect detection method based on deep learning is gradually used in the defect recognition of MPI [11].…”

Section: Introductionmentioning

confidence: 99%

“…The comparison demonstrates that the semantic segmentation-based approach ensures the highest defect localization accuracy while being the fastest for realtime applications. Only the defect images can be used to train the network, thus requiring a large training dataset [11]. Japanese scholar Tohru Kamiya et al improved the U-net for fluorescent MPI image features by adding a convolutional layer to segment and classify little defects without losing features.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Defect detection algorithm for magnetic particle inspection of aviation ferromagnetic parts based on improved DeepLabv3+

Zeng

Zhang

Wang³

et al. 2023

Meas. Sci. Technol.

View full text Add to dashboard Cite

Non-destructive testing (NDT) of magnetic materials such as aviation parts is an indispensable part of the civil aviation maintenance industry. The NDT of such metal materials often uses magnetic particle inspection (MPI) technology. This paper proposes an improved DeepLabv3+ semantic segmentation algorithm for automatic defect detection of aviation ferromagnetic parts after MPI. In the network structure, lightweight MobileNetV2 is the backbone feature extraction network. The Dense Atrous Spatial Pyramid Pooling (DenseASPP) structure is used to strengthen feature extraction. The influence of three different DenseASPP structures on the recognition effect is compared in the experiment. At the same time, the decoder is further optimized. The experimental results show that the Ours-DeepLabv3+ network model can effectively for automatic defect detection of aviation ferromagnetic parts after MPI. The Precision, Recall, F1-score, and IOU are 81.64%, 83.12%, 82.37%, and 71.23%, respectively, which are 7.48%, 5.45%, 6.50%, and 10.1% higher than the original DeepLabv3+, and defect detail segmentation is more accurate. Compared with other semantic segmentation algorithms, this method can effectively improve the accuracy of defect detection of aviation ferromagnetic parts and meet the requirements of defect detection.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Defect detection algorithm for magnetic particle inspection of aviation ferromagnetic parts based on improved DeepLabv3+

Zeng

Zhang

Wang³

et al. 2023

Meas. Sci. Technol.

View full text Add to dashboard Cite

show abstract

“…Therefore, it is of great engineering significance to improve the crack detection capability of bearing rings and promote their quality [2]. At present, the defect detection methods for bearing rings include magnetic particle testing [3,4], eddy current testing [5,6], magnetic flux leakage (MFL) testing, ultrasonic testing [7,8], and machine vision testing [9][10][11]. MFL testing is widely used to evaluate various ferromagnetic materials.…”

Section: Introductionmentioning

confidence: 99%

Influence of curvature radius on the axial crack signal of the magnetic flux leakage detection for tapered roller bearing rings

Yang

Peng²,

Qiu³

et al. 2023

Front. Phys.

View full text Add to dashboard Cite

Tapered roller bearings are widely used in heavy machinery, railway transportation, aviation, and other fields. Their quality and reliability are related to the operational safety of mechanical equipment. In the axial crack magnetic flux leakage (MFL) detection of the bearing ring, the MFL signals obtained by the sensor from different curvature radius of the surface are inconsistent, affecting the detection accuracy of cracks and subsequent quantitative analysis. In order to address the above problems, the finite element simulation is performed to analyze the influence of the surface curvature radius of the bearing ring on the magnetic field distribution inside the workpiece and the MFL signal in the circumferential magnetization. Through the parallel magnetic circuit, the difference in curvature radius is identified as the basic reason for non-uniform magnetization. On this basis, the compensation method based on the normalization of surface magnetization is proposed. Furthermore, the effectiveness of the compensation method is verified by experiments. The relative change in the amplitude of the crack MFL signal is reduced from 30% to 5%.

show abstract

“…General testing technologies need to sample research materials and conduct a destructive test, which is not conducive to the maintenance of the automated machine. Non-destructive testing technologies, including magnetic particle testing [ 5 , 6 , 7 ], radiographic testing [ 8 ], and eddy current testing [ 9 ], can detect the health condition without damaging the tested materials. However, some common non-destructive testing technologies also face problems such as the requirement to keep the tested equipment static, complex analysis processes, and so on.…”

Section: Introductionmentioning

confidence: 99%

Non-Destructive Testing for Cavity Damages in Automated Machines Based on Acoustic Emission Tomography

Dong

Pei

2022

Sensors

View full text Add to dashboard Cite

Damage detection is important for the maintenance of automated machines. General non-destructive testing techniques require static equipment and complex analysis processes, which restricts the maintenance of automated machines. Therefore, this paper proposes an acoustic emission (AE) tomography method for detecting cavity damage in automated machines, combining the fast sweeping method (FSM) and the limited-memory Broyden–Fletcher–Goldfarb–Shanno (L-BFGS) method. This approach overcomes the limitations of real-time AE detection for cavity damage in continuous and homogeneous materials. The proposed method has been applied in numerical and laboratory experiments to validate its feasibility. The results show that the inversed low-velocity regions correspond to the actual cavity regions, and the sources of cavity damage can be effectively detected. This paper provides a new perspective for AE testing technologies, and also lays the foundation for other non-destructive testing techniques, in terms of cavity damage detection.

show abstract

Automated vision system for magnetic particle inspection of crankshafts using convolutional neural networks

Cited by 25 publications

References 39 publications

Defect detection algorithm for magnetic particle inspection of aviation ferromagnetic parts based on improved DeepLabv3+

Defect detection algorithm for magnetic particle inspection of aviation ferromagnetic parts based on improved DeepLabv3+

Influence of curvature radius on the axial crack signal of the magnetic flux leakage detection for tapered roller bearing rings

Non-Destructive Testing for Cavity Damages in Automated Machines Based on Acoustic Emission Tomography

Contact Info

Product

Resources

About