Open Access
Add to BibSonomy
Abstract
This paper examines the characteristics of magneto-optical images of weld defects under alternating magnetic field excitation. Weld defects such as non-penetration, surface cracks and sub-surface cracks were detected by a magneto-optical imaging method. Magneto-optical imaging nondestructive testing experiments under alternating magnetic field excitation were carried out to detect the weld defects. Image processing methods which include contrast enhancement of original image, fused image, contrast enhancement of fused image were applied to extract the defect information of the magneto-optical images. What’s more, the difference among the magneto-optical images of weld defects was obtained by contrast analysis. Experimental results show that non-penetration welding images possess significant differences in brightness and darkness, and this difference in cracks is smaller than non-penetrating ones. Under the same excitation conditions, the leakage flux of welds with non-penetration is stronger than that of weld cracks.
© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Effective nondestructive testing must be carried out to ensure the quality of welding products, in which the weld cracks detection and non-penetration detection are of the utmost concern [1]. Radiographic testing [2,3], ultrasonic testing [4,5], osmotic testing, magnetic particle testing [6], magnetic flux leakage testing and eddy current testing [7,8] are the main conventional methods for weld defects detection. Radiographic testing requires large equipment and costs too much. What’s more, it does much damage to human body. Ultrasonic testing requires a coupling agent and has a high demand for the operator. Osmotic testing is susceptible to human factors and is poor in the detection of samples with surface moisture or film. Magnetic particle testing has the drawback that the weld to be tested must be paramagnetic material and be pre-treated as well. Magnetic flux leakage detection and eddy current testing are only suitable for the surface and sub-surface detection of conductive materials. And it is difficult to acquire the type, shape and size of weld defects.
Nowadays, magneto-optical imaging detection of weld defects is mainly based on constant magnetic field excitation, which is applied to micro-gap seam tracking and weld nondestructive testing. High contrast magneto-optical images are obtained by using electromagnet, which is applied to the weld inspection [9,10]. Magneto-optical imaging(MOI) technology has been approved for the crack inspection in aging air craft [11,12]. This technique is pinpoint accurate and high efficient in measuring and imaging the magnetic field distribution [13]. As a novel nondestructive testing method, it equips the inspector an ability to generate images quickly in aviation material surface and sub-surface fine cracks in real-time [14]. The image of the magneto-optical imaging nondestructive testing technique for weld defects is basically obtained under the constant magnetic field excitation. And there is a shortage of missing weld information due to the easy saturation. In view of the above problems, the non-penetration, surface cracks and sub-surface cracks of different gaps are weld defects obtained by butt joint welding. Under the alternating magnetic field excitation, these defects were subjected to magneto-optical imaging nondestructive testing to analyze the imaging characteristics. And this article is an expanded version of the IEEE IST 2017 conference paper [15]. The characteristics of magnetic field distribution in weld cracks and non-penetration are summarized to optimize the nondestructive testing method of magneto-optical imaging under alternating magnetic excitation.
2. Principle of alternating magneto-optical imaging
Based on the Faraday magnetism [16], the magneto-optical imaging working principle of detecting the weld is shown in Fig. 1. Humans have long studied the propagation of light, and the propagation of light in magnetic fields is used to detect defects in the weld. Linearly polarized light which is altered from the monochromatic high power light passing through the polarizer is detected by the analyzer after passing through the magneto-optical film and the reflecting lens in turn, which is then received by the imaging element to form a light intensity map. The polarization direction of the linearly polarized light rotates at a certain angle in the magnetic field, and the rotation angle θ is proportional to the distance between the light passing through the medium and the magnetic induction intensity component B in the propagation direction of the light in the medium. This rotation angle Ө can be expressed as
where V (in radians per tesla per meter) is the Verdet constant for material. The corresponding MO images of Ii (i = 0, 1, 2) are shown on the top of Fig. 1. Light rotates clockwise when the electromagnet is in the N pole field, which is corresponding to the bright part of the magneto-optical image, while the light rotates anticlockwise when the electromagnet is the S pole field, which corresponds to the dark part of the magneto-optical image [17].During the test, AC electromagnet powered by a sinusoidal voltage of 50 Hz is selected as the excitation source. The sampling frequency of the magneto-optical sensor is 75 frame/s, and its lift-off degree is set to 0.5 mm. With the change of the sampling time point, the magnetic field and the magnetic field intensity of the excitation field change accordingly, and the collecting magneto-optical image changes as well. The ideal sampling point distribution is shown in Fig. 2.
As seen from Fig. 2, the domains were changing along with the voltage sine wave, thus a single MO image can be captured in every 13.3 ms(millisecond). The relation between frame rate and voltage frequency is illustrated as below in which three dynamic images of frame 1, frame 2 and frame 3 were presented, arranged as a, b and c, respectively. The information obtained from three continuous magneto-optical images with COMS sensor is weld seam information under different magnetic induction and magnetic field directions [18]. In the magneto-optical images collected at different sampling start points, the magnetic field direction and the magnetic field size of each successive three-frame image are different.
Alternating magnetic field is selected according to the hysteresis loop characteristic of the magneto-optical thin film which is stipulated that the induced magnetic field of the weld to be measured can’t be greater than 2.5mT. If the induced magnetic field exceeds 2.5 mT, the magneto-optical image may be too bright or too dark due to saturation, and the important weld information is lost. Due to the small magnetic induction of the excitation field and the presence of hysteresis, the weld can’t undergo a complete hysteresis loop during excitation. As the intensity of the excitation field is small and its direction is changing constantly, the workpiece forms some new and small curves in the middle of the original hysteresis loop during the excitation. The selection of the sampling frequency of the magneto-optical imaging sensor determines the formation characteristics of the new hysteresis loop during the welding process.
3. Magneto-optical imaging test of weld defects
The steel plate in the experiment was 100 mm × 50 mm × 2 mm (length × width × thickness). Steel plate spot welding is used to simulate the weld defect. Weld defects including non-penetration, surface cracks and sub-surface cracks are simulated by a laser butt joint welding, and their gap widths were about 0.01 mm, 0.05 mm and 0.1 mm respectively. Nondestructive testing system is shown in Fig. 3.
It was found that the magneto-optical image was the best effect when the excitation voltage was 200 volts. The magnetic flux decreases (or increases) because the weld cross-sectional area becomes smaller (or becomes larger) when there is a defect, which makes the magnetic field lines squeezed at the defect and thus produces a leakage magnetic field [19,20]. The frame rate of magneto-optical sensor was 75 fps and image resolution was 400pixel × 400pixel. The magneto-optical imaging sensor reflects morphological characteristics of the defect by examining small changes in magnetic field. Different weld defects in the excitation field have different leakage magnetic fields, and the gray value of obtained magneto-optical image reflects the strength of leakage magnetic field.
3.1 Magneto-optical Images of Non-penetration Welds
Non-penetration is simulated by spot welding on both ends of the butt plate. The image obtained under the excitation of the alternating magnetic field, such as the cross-section of non-penetration, the physical map of the region of interest and the corresponding adjacent three-frame magneto-optical image, is shown in Table 1. It contains approximate defect gap size(ADGS), schematic cross section of weld(SCSW), partial magnification of object(PMO), first frame(F1), second frame(F2), third frame(F3), contrast enhancement of original image(CEOI), fused image(FI), and contrast enhancement of fused image(CEFI).
Table 1. Magneto-optic Images of Non-penetration Welds under Alternating Excitation
As seen from Table 1, there is a large difference between the magneto-optical images of the defect-free weldment and non-penetration weldment. The magneto-optical image of defect-free weldment has a uniform light intensity distribution. The magneto-optical images of the non-penetration have a clear dividing line, in which one part of the image has a significant difference between the two sides of the dividing line, while the other part of the image boundary is dark or bright on both sides.
In the case of nondestructive testing of magneto-optical imaging under alternating magnetic field excitation, the strength variation of the magnetic field stabilizes near a value due to the influence of hysteresis. Magneto-optical image obtained by magneto-optical imaging sensor does not exhibit a large difference in light intensity because the sampling start point is different. As the image is either too bright or too dark, it is not conducive for the extraction of the welding features. So the magneto-optical image that can demonstrate the difference between the bright and the dark is selected in Table 1. For example, the second frame with a gap about 0.1 mm, the third frame with a gap about 0.05 mm and the second frame with a gap about 0.01 mm are chosen as testing sample. These selected magneto-optical images are subjected to image processing to obtain a Contrast enhancement image and the gray value of image in column.
As shown in Table 1, the three dynamic MO images from frame1 (F1), frame 2 (F2) and frame 3 (F3) were acquired in two alternating magnetic field periods. There is a large difference between the magneto-optical images of defect-free welds and non-penetration welds. A graph is obtained from each set of images in the Table 1, which has a distinct transition zone. The gray value of that image which has been marked with a red box in column is extracted. Image fusion technology is used in many fields [21–24]. And the three frames were regarded as one fused image by using weighted average method which belongs to pixel-level fusion [25,26]. The fused image (Ff ) can be obtained as follows:
where a, b and c are the weight of each frame, the total weight is equal to 1. Therefore an ideal couple in which a, b and c were set as 0.1, 0.7, and 0.2 was used in this experiment. The bright and dark distribution of the non-penetration continuous three-frame magneto-optical image is very similar, so we can use the same values as the weighted coefficients of corresponding images. The most suitable weighting coefficients are determined by the quality of fusion image. It should ensure that the fusion image is not saturated and the defect distribution is obvious.The contrast enhancement of original images, fused images and contrast enhancement of fused images are shown in Table 1. It’s easy to find that the original magneto-optical image with defect is lighter than the weighted image. The magneto-optical images of the welds without defects are substantially unchanged before and after image processing. The gray value of image in column is extracted, and the results are shown in Fig. 4. After image fusion, the defects of the images are obvious.
Fig. 4 Gray value in column. (a) Image without defective weld. (b)Original image of non-penetration in column. (c)Fused image of non-penetration in column
In the case of non-penetration, the gray value of the original image in the column is linearly distributed. The linear area of the gap of about 0.1 mm is the largest, and the gap about 0.01 mm is the smallest. This indicates that the larger the gap is, the larger the bright area of the magneto-optical image is, which proves that the larger the gap is, the more intense the leakage magnetic field is. The gray value with no defect is more concentrated. It can be seen that the values of 50 to 300 in the grayscale curve in Fig. 4(b) are constant. Single-frame images under alternating excitation are easy to saturate as magneto-optical images obtained under constant field excitation. Grayscale curve represents the strength of the leakage magnetic field, which will appear constant value area after saturation. It can be concluded that single-frame magneto-optical images can be easily saturated and thus lose the important weld information. The reason that the curve is not smooth is that the original image is not fully filtered since it contains the noise and texture features of the steel plate.
It can be seen from Fig. 4(c) that the gray value curve of fused image in column has no linear distribution area. Although the left curve is relatively gentle, its sloping inclination maintains. And the distribution of curves is the same as that of non-penetrating magnetic flux leakage. This is consistent with the distribution of gray value in unsaturated state. Comparing Fig. 4(b) to 4(c), it can be seen that weld information of single magneto-optical images will be lost due to saturation phenomenon. So it can be summarized that image fusion technology can reduce the image saturation of the weld information loss.
3.2 Magneto-optical Images of Surface Cracks
Surface cracks of the weld is simulated by pulsed laser welding on the upper and lower surfaces of the abutting steel plate, and it must be ensured that weld on the upper surface is shorter than the weld on the lower surface. The image obtained under the excitation of the alternating magnetic field, such as the cross-section of surface crack, the physical map of the region of interest and the corresponding adjacent three-frame magneto-optical image, is shown in Table 2. Furthermore, the three frames were regarded as one fused image by using weighted average method. Therefore, values of a, b, and c of fused image are set to 0.1, −0.1 and 1.0 in the gap of 0.01 mm; and in the gap of 0.05 mm, the value of a, b, and c are 0.5, −0.1 and 0.6 respectively; and in the gap of 0.1 mm, the value of a, b, and c are −0.1, 0.3 and 0.8, respectively.
Table 2. Magneto-optical Images of Surface Cracks in Alternating Excitation
It can be seen from Table 2 that the magneto-optical images of weld surface cracks change obviously at the defects, and the images appear as a stripe-like boundary. Magneto-optical image of the base material area on both sides of the weld is basically the same, and the light intensity distribution is uniform. Surface cracks are slightly delimited with the base metal zone and the weld zone. As the gap of cracks increases, the difference between crack and other regions of the magneto-optical image is more pronounced. The magneto-optical image whose crack areas are darker than the base metal area is selected in Table 2. For example, the third frame with a gap about 0.1 mm, the second frame with a gap about 0.05 mm, and the third frame with a gap about 0.01 mm are chosen as test sample. These selected magneto-optical images are subjected to image processing to obtain a contrast enhancement image and the gray value of image in column.
The contrast enhancement of original images, fused images and contrast enhancement of fused images are shown in Table 2. It can be seen that the original magneto-optical images and the fused magneto-optical images of defects are more obvious, and fused images have no obvious advantages than the original. The gray value of image which has been marked with a red box in column is extracted. The gray value in column is shown in Fig. 5.
As shown in Fig. 5(a), it can be seen from the right side of the curve that the gap of 0.01 mm tends to be saturated, the 0.05mm gap image does not appear saturated, and the image saturation of the 0.1mm gap is obvious. This indicates that under the same experimental conditions, the larger the gap is, the easier the magneto-optical image will saturate. From Fig. 5(b), it can be seen that the saturation phenomenon of magneto-optical image is improved by image fusion, and the gray value curve has the boundary of obvious transition zone. This shows that image fusion technology overcomes the shortcomings of image saturation and prevents loss of weld information effectively, which improves the efficiency and accuracy of weld defect detection.
3.3 Magneto-optical Images of Sub-surface Cracks in Welds
Sub-surface crack of the weld is simulated by pulsed laser welding on upper and lower surfaces of the butt plate, and the weld is not welded in the middle. In Table 3, there are images obtained under the excitation of the alternating magnetic field, such as the cross-section of sub-surface crack, the physical map of the region of interest and the corresponding adjacent three-frame magneto-optical image. Therefore, the values of a, b, and c of fused image are set to 0.1, −0.1 and 1.0 in the gap of 0.01 mm; and in the gray of 0.05 mm, the values of a, b, c are 0.1,1.2 and −0.3 respectively; and in the gap of 0.1 mm, the value of a, b, c are 0.1, −0.1 and 1.0, respectively.
Table 3. Magneto-optical Images of Sub-surface Cracks in Weld under Alternating Magnetic field Excitation
It can be seen from Table 3 that the magneto-optical image of sub-surface crack in the weld has a band-like boundary difference at the light intensity of the defect. The light intensity of magneto-optical image in the base material area on both sides of the weld is uniform and there is no obvious difference in brightness. The difference in light intensity at the crack of magneto-optical image becomes larger accordingly as the gap of sub-surface becomes larger.
Magneto-optical images whose sub-surface crack regions are darker than the parent metal region are selected. For example, the first frame with a gap about 0.1 mm, the second frame with a gap about 0.05 mm and the first frame with a gap about 0.01 mm are selected. These selected magneto-optical images are subjected to image processing to obtain an enhancement image and its gray value. The results shown in Table 3 and Figs. 6. Width d1 and d2 are defined as the distance between two peak inflection points. The inflection point of gray value is that the magnetic force line overflows at the defect boundary and the magnetic induction intensity changes.
As seen from Table 3, when the contrast of original image was increased, the defect is more obvious, and the defect of fused image is clearer than the original. By observing the magneto-optical image of the original image and the fusion image, it can be found that compared with the defect boundary of the gap with 0.05 mm and 0.01mm, that of the gap with 0.1mm is relatively unclear. The reason is that the larger the gap is, the stronger the leakage magnetic field (that is, the overflow of magnetic field) is, and the magneto-optic image is not saturated. This leads to unclear boundaries. The gray value of fused image is related to the curve of 0.05mm gap, has a more obvious transition zone. The yellow line in the Figs. 6(a) and 6(b) indicates where leakage field changes the direction of magnetic field. It can be seen from the Figs. 6(a) and 6(b) that d1 is larger than d2, and every 102 pixels represents 1 mm in the magneto-optical image, which indicates that the fused image is closer to the size of the defect before fusing. It can be found that the measured value is larger than the actual value, and the reason for this phenomenon is the lift-off degree of the sensor. The larger the lift-off degree, the larger the defect size measured. The laboratory has also studied the influence of lift-off degree on the test results, but there is no quantitative analysis. Thank you for your question, which will also be a point of my later research. Observing the grayscale value curve of 0.01mm and 0.1mm in Fig. 6(a), it can be found that the turning point on the left is not obvious on the right, the fused image is clearer than the original image, so it is easy to make errors as judging the defect contour. This is the reason that d2 is smaller than d1. In the gray value graph, 0.05 mm gap curve is quite different. By checking the experimental process, it is found that the excitation device has been moved so that the experimental conditions are not the same as before.
4. Contrast analysis of experimental results
From Table 1 to Table 3, it can be found that the images of different defects have the same rules, that is the continuous three-frame magneto-optical images of different gap defects including the bright and dark half-graphs, full-bright and all-dark graphs. This proves that the type and size of defects do not affect the distribution of bright and dark regions of magneto-optical image when the defect detection is performed, which only changes the brightness of the image. This will play a guiding role in analyzing the magneto-optical images of weld defects in the future.
From Figs. 4(a)-4(c) to Figs. 6(a)-6(b), it is clearly observed that that the magneto-optical image obtained by alternating excitation can avoid the loss of weld information caused by image saturation after image fusion. The results of extracting the extreme value of the gray value curve are shown in Table 4, which contains defects(D), gap of defects(GD), image (I), extraction of the extreme value of gray value curve(EEG), no defects(ND), non-penetration(NP), surface cracks(SC), and sub-surface cracks(SSC). As can be seen from the table, the maximum gray value of the fused non-penetrating images is less than 255, and the difference between the maximum and the minimum value is close to 100. After image-weighted fusion processing, the loss of weld information caused by the saturation of the image is appropriately reduced. For surface cracks, the difference between the maximum and minimum values of the fused image is less than the unfused image. Combining this table with Figs. 6, it can be found that the difference of pixel value of the fused image is not necessarily larger than that of the unfused image but still improves the saturation of the image. As for sub-surface cracks, from Table 4, it can be found that the difference between the maximum and minimum values of the fused image is substantially greater than the unfused image.
Table 4. Extract the extreme value of gray value curve
In magneto-optical imaging experiments, the magnetic poles distributed in the non-penetrating weld image are consistent. It is proved that the welds form a new stable hysteresis loop in magneto-optical imaging nondestructive testing experiment under alternating excitation, and the obtained magneto-optical images do not vary according to the different starting point of sampling.
In the experiment, the magneto-optical image of non-penetration welds has a obliquely upward tendency, and the leakage magnetic field of surface cracks and sub-surface cracks is not exactly the same as the actual crack shape. The reason for this phenomenon is related to the position of magnetic pole of exciter. If the excitation field can’t guarantee a uniform distribution of the transient field, the leakage magnetic field distribution will be different.
Compared with previous work which is the characteristics of magneto optical nondestructive testing under constant magnetic field and image processing or pattern recognition of magneto-optic image obtained under the excitation of alternating magnetic field, this article focuses on the characteristics of weld leakage field under alternating magnetic field, which helps to optimize the nondestructive testing of magneto-optic imaging. The results of three experiments show that the larger the gap is, the greater the leakage magnetic field is. Under the same experimental conditions, the saturation of magneto-optical image of non-penetration is larger than that of surface crack, and the saturation of sub-surface crack is the smallest.
5. Conclusion
This paper introduces the test principle of nondestructive testing of magneto-optical imaging under alternating magnetic field excitation. Weld defects including non-penetration, surface crack and sub-surface cracks, are subjected to magneto-optical imaging nondestructive testing experiments under the 50 Hz alternating electromagnetic field excitation, in which the gap widths are about 0.01 mm, 0.05 mm and 0.1 mm, respectively. Experimental results show that there is a clear linear boundary line with non-penetrating magneto-optical image, and the larger the gap is, the larger the bright area of magneto-optical image is. Nondestructive testing of magneto-optical imaging under alternating magnetic field excitation can effectively improve the shortcomings of magneto-optical image saturation under constant magnetic field excitation. The fused image of subsurface defects under alternating excitation is clearer than unfixed images. Under the same experimental conditions, the effect of single-frame magneto-optical map is worse than that of surface defects. For the magneto-optical image of the weld cracks with a band-like light and dark transition zone, the larger the gap of crack is, the larger the transition zone is and the stronger the leakage magnetic field is. Under the same excitation conditions, the non-penetration magnetic induction intensity is the largest, and the sub-surface crack is the weakest. As for the nondestructive testing of magneto-optical imaging under alternating excitation, due to hysteresis, the size of the weld-induced magnetic field changes around a fixed value. Therefore, the light intensity obtained by sampling the magneto-optical imaging sensor does not differ in the size of the magnetic induction due to the difference in the starting point of the sampling.
Funding
National Natural Science Foundation of China (NSFC) (51675104); The Science and Technology Planning Project of Guangzhou, China (201510010089); The Science and Technology Planning Public Project of Guangdong Province, China (2016A010102015).
References and links
1. A. Ancona, V. Spagnolo, P. M. Lugarà, and M. Ferrara, “Optical Sensor for real-time Monitoring of CO(2) Laser Welding Process,” Appl. Opt. 40(33), 6019–6025 (2001). [CrossRef] [PubMed]
2. C. Yu, B. Chang, S. Wang, and Q. Wang, “Design principles of a novel X-ray imaging system,” Chin. Opt. Lett. 4(3), 158–159 (2006).
3. K. Xu, Z. Zhang, Q. Yu, and Z. Wen, “Field-Effect Electroluminescence Spectraof, Reverse-Biased PN Junctions in Silicon Device for Microdisplay,” J. Disp. Technol. 12(2), 115–121 (2016).
4. G. Sun, Z. Zhou, X. Chen, and J. Wang, “Ultrasonic characterization of delamination in aeronautical composites using noncontact laser generation and detection,” Appl. Opt. 52(26), 6481–6486 (2013). [CrossRef] [PubMed]
5. Q. Wu and Y. Okabe, “High-sensitivity ultrasonic phase-shifted fiber Bragg grating balanced sensing system,” Opt. Express 20(27), 28353–28362 (2012). [CrossRef] [PubMed]
6. Z. F. Zhou and Y. H. Cheng, “Magneto-optic microscope for visually detecting subsurface defects,” Appl. Opt. 47(19), 3463–3466 (2008). [CrossRef] [PubMed]
7. W. Swiderski, “Detecting defects in marine structures by using eddy current infrared thermography,” Appl. Opt. 55(34), D17–D21 (2016). [CrossRef] [PubMed]
8. J. M. Prince and B. P. Hildebrand, “Low-frequency electromagnetic (eddy-current) holography for imaging in conductors,” Appl. Opt. 32(26), 4960–4971 (1993). [CrossRef] [PubMed]
9. X. Gao, C. Lan, D. You, G. Li, and N. Zhang, “Weldment Nondestructive Testing Using Magneto-optical Imaging Induced by Alternating Magnetic Field,” Journal of Nondestructive Evaluation 36(3), 55 (2017). [CrossRef]
10. X. Gao, Y. Chen, D. You, Z. Xiao, and X. Chen, “Detection of micro gap weld joint by using magneto-optical imaging and Kalman filtering compensated with RBF neural network,” Mech. Syst. Signal Process. 84, 570–583 (2017). [CrossRef]
11. G. L. Fitzpatrick, D. K. Thome, R. L. Skaugset, E. Y. C. Shih, and W. C. L. Shih, “Magneto-optic/eddy current imaging of aging aircraft: a new NDI technique,” Mater. Eval. 51(12), 1402–1407 (1993).
12. Z. Zeng, X. Liu, Y. Deng, L. Udpa, L. Xuan, W. C. L. Shih, and G. L. Fitzpatrick, “A Parametric Study of Magneto-Optic Imaging Using Finite-Element Analysis Applied to Aircraft Rivet Site Inspection,” IEEE Trans. Magn. 42(11), 3737–3744 (2006). [CrossRef]
13. K. Xu, Z. Zhang, Q. Yu, and Z. Wen, “Monolithically integrated Si gate-controlled light-emitting device: science and properties,” J. Disp. Technol. 12(2), 115–121 (2016).
14. Y. Deng, X. Liu, and L. Udpa, “Magneto-Optic Imaging for Aircraft Skins Inspection: A Probability of Detection Study of Simulated and Experimental Image Data,” IEEE Trans. Reliab. 61(4), 1–8 (2011).
15. N. Ma, X. Gao, Y. Song, and N. Zhang, “Magneto-optical imaging characteristics of weld defects under alternating magnetic field excitation,” in 2017 IEEE International Conference on Imaging Systems and Techniques (IST) (2017), pp. 394–398. [CrossRef]
16. M. Koschny and M. Lindner, “Magneto-optical sensors accurately analyze magnetic field distribution of magnetic materials,” Adv. Mater. Process. 170(2), 13–16 (2012).
17. X. Gao, G. Huang, D. You, C. Lan, and N. Zhang, “Magneto-optical imaging deviation model of micro-gap weld joint,” J. Manuf. Syst. 42, 82–92 (2017). [CrossRef]
18. K. Xu, “Integrated silicon directly modulated light source using p-well in standard CMOS technology,” IEEE Sens. J. 16(16), 6184–6191 (2016). [CrossRef]
19. H. Richert, H. Schmidt, S. Lindner, M. Lindner, B. Wenzel, R. Holzhey, and R. Schäfer, “Dynamic magneto-optical imaging of domains in grain oriented electrical steel,” Steel Res. 87(2), 232–240 (2016). [CrossRef]
20. M. Le, J. Lee, and T. Shoji, “A simulation of magneto-optical eddy current imaging,” NDT Int. 44(8), 783–788 (2011). [CrossRef]
21. L. Chen, J. Li, and C. L. Chen, “Regional multifocus image fusion using sparse representation,” Opt. Express 21(4), 5182–5197 (2013). [CrossRef] [PubMed]
22. X. Bai, F. Zhou, and B. Xue, “Fusion of infrared and visual images through region extraction by using multi scale center-surround top-hat transform,” Opt. Express 19(9), 8444–8457 (2011). [CrossRef] [PubMed]
23. L. Guo, M. Dai, and M. Zhu, “Multifocus color image fusion based on quaternion curvelet transform,” Opt. Express 20(17), 18846–18860 (2012). [CrossRef] [PubMed]
24. Q. Guihong, Z. Dali, and Y. Pingfan, “Medical image fusion by wavelet transform modulus maxima,” Opt. Express 9(4), 184–190 (2001). [CrossRef] [PubMed]
25. M. Nooshyar, M. Abdipour, and M. Khajuee, “Multi-focus Image Fusion for Visual Sensor Networks in Wavelet Domain,” Comput. Electr. Eng. 37(5), 789–797 (2011). [CrossRef]
26. O. Prakash, A. Kumar, and A. Khare, “Pixel-level image fusion scheme based on steerable pyramid wavelet transform using absolute maximum selection fusion rule,” in International Conference on Issues and Challenges in Intelligent Computing Techniques (IEEE, 2014), pp. 765–770. [CrossRef]
