-
在空域上图像增强主要是针对图像像素而言,通过改善图像的对比度、灰度来增强图像[14-15]。直方图均衡化是图像增强的基本方式之一[16-17],主要是利用灰度变换自动调节图像对比度。
其方法是:首先将彩色图像转换为灰度图像,灰度图像有256个灰度级,假设转换后的图像灰度值为r,第k级的灰度值为rk,直方图均衡化后的图像中,第k级的灰度值为sk,则原直方图灰度级分布概率Pr(rk)可表示为:
$ P_{r}\left(r_{k}\right)=\frac{n_{k}}{n} $
(1) 式中,n是图像中像素总和,nk是第k级灰度值为rk的像素个数。计算直方图概率累计函数Ps(sk):
$ {P_s}\left( {{s_k}} \right) = \sum\limits_{j = 0}^k {\frac{{{n_j}}}{n}} $
(2) 式中,nj是第j级灰度值为sj的像素个数。通过概率累计率与原图像灰度级进行取整拓展, 得到均衡化后第k级的灰度值为:
$ s_{k}=\operatorname{int}\left\{\left[\max \left(r_{k}\right)-\min \left(r_{k}\right)\right] \cdot P_{s}\left(s_{k}\right)+0.5\right\} $
(3) -
磁极0°时,对磁光图像通过直方图均衡化进行图像增强,直方图均衡化后的灰度图及灰度变化曲线如图 7所示。
由图 7中水平方向与垂直方向上的灰度变化曲线可知,水平方向上的裂纹两侧灰度值发生了剧烈变化,且纵向变化区域200pixel~250pixel之间存在明显的明暗分界线。同样对磁极为45°,90°,135°时的磁光图像进行直方图均衡化处理,获得灰度图像和水平垂直方向上的灰度变化曲线,如图 8所示。
由图 7和图 8可知,当磁极角度从0°~90°旋转时,励磁装置改变励磁角度时,垂直方向上的缺陷会越来越明显。当角度到达90°时,垂直方向上的缺陷最明显,磁光图像上缺陷的分界线位于像素150pixel~240pixel之间,且垂直裂纹两侧的灰度值发生剧变,而此时水平方向上的缺陷信息已经很难识别到。当励磁装置改变角度至135°时,水平方向上的缺陷重新显现,从灰度变化曲线中可以明显看出,两个方向上的裂纹交界处均有灰度数值剧变,都出现在200pixel附近,但跨度较大。
通过恒定磁场励磁下多角度励磁, 验证了垂直通过裂纹的磁场产生的漏磁能最有效地使得磁光传感器成像,同时切向磁场也能使缺陷暴露,但变化区域不明显。
-
交变励磁下所采集的磁光图像呈现3帧一个周期的变化。因此在每个角度励磁稳定后选取3帧连续的磁光图像,这3帧是不同励磁强度下获得的磁光图像。由于磁极的分布位置一样,不同的只是磁场强度,故对3帧图像进行图像融合[12],将3帧磁光图像上的特征融合呈现在一幅图像上[18-20]。融合后的图像如图 9所示。
对融合后的灰度图像进行直方图均衡化处理。直方图均衡化灰度图和灰度曲线变化图如图 10、图 11、图 12和图 13所示。
由图 10可知,当励磁角度为0°时,可以得到明显的横向缺陷图像信息,从水平垂直灰度分布曲线上可以看到,裂纹的分界线位于垂直像素195pixel~220pixel处,两侧的图像灰度值变化明显; 但纵向裂纹信息明显不足,从分布曲线上无法表现出来。
由图 11可知,当励磁角度为45°时,可以得到横向纵向两个角度的缺陷信息。从水平垂直灰度分布曲线上可以看到,横纵两个方向上的均能明显成像,裂纹的分界线位于200pixel两侧,均能呈现出明显的边界。但与0°励磁相比,横向裂纹的位置出现了轻微的上偏移。图中左边出现了两个边界,下边界靠近磁极,出现边界线向磁极放线弯曲的现象。此时纵向边界出现在180pixel~190piexl处。
由图 12可知,当励磁角度是90°时,可以得到纵向裂纹明显的缺陷信息,分界线位于190pixel~200pixel两侧,同时横向裂纹也有所显现,但过渡带太大,不能明显得到边界位置信息。
由图 13可知, 当励磁角度为135°时,重新可以得到横向纵向两个角度的缺陷信息,从均衡化后的灰度分布曲线上可以看到,横纵两个方向上的均能明显成像,横向裂纹分界线位于190pixel~200pixel两侧,纵向裂纹分界线位于195pixel~205pixel均能呈现出明显的边界。与45°角励磁相比较,此时可以看到图像中上下两个边界线都向磁极所在区域发生弯曲。
激光焊接缺陷多向磁场激励下磁光成像检测
Magneto-optical imaging detection of laser welding defects under multi-directional magnetic field excitation
-
摘要: 为了研究交叉焊缝缺陷的磁光图像特征,对中碳钢板进行十字对接激光焊,获得同时具有横向和纵向焊后裂纹的试验样本,同时采用恒定磁场和交变磁场对试验样本进行多角度励磁, 用磁光成像传感器采集不同角度磁场激励下焊接缺陷处的磁光图像,并对磁光图像的缺陷特征进行分析。结果表明,多向磁场激励下的磁光成像技术能明显检测出多角度的焊接缺陷,且能有效避免曲线裂纹在焊接缺陷检测中的漏检现象; 同时,交变磁场激励下, 同一裂纹成像的分辨率提高了40pixel~50pixel,能更精确地定位缺陷位置。此研究为提高焊接缺陷的检出率提供了依据。Abstract: In order to study the magneto-optical image(MOI) characteristics of cross-weld defects, cross butt laser welding was performed on the medium carbon steel plate to obtain test samples with both transverse and longitudinal cracks. The test sample was excited by constant magnetic field and alternating magnetic field. Magneto-optical imaging sensor was used to collect magneto-optical images of welding defects under magnetic excitation from different angles. The defect characteristics of magneto-optical image were then analyzed. The results show that the magneto-optical imaging technology under multi-direction excitation can obviously detect multi-angle welding defects, and can effectively avoid the missing detection phenomenon of curve cracks in welding defect detection. The resolution of the same crack image is increased by 40pixel~50pixel under the excitation of the alternating magnetic field, which can locate the defect more accurately. This study provides a basis for improving the detection rate of welding defects.
-
Figure 8. Gray images and gray curve images of cross crack under different angle excitation
a—histogram equalized image with excitation angle 45° b—the curves of gray level of 45° c—histogram equalized image with excitation angle 90° d—the curves of gray level of 90° e—histogram equalized image with excitation angle 135° f—the curves of gray level of 135°
-
[1] YOU D Y, GAO X D. Studies and prospects of laser welding techno-logy[J]. Welding Technology, 2008, 37(4):5-9(in Chinese). [2] KOSCHNY M, LINDNER M. Magneto-optical sensors accurately analyze magnetic field distribution of magnetic materials[J]. Advanced Materials & Processes, 2012, 170(2):13-16. [3] ZENG H M, GAO X D, ZHANG N F. Nondestructive testing methods of welding crack[J]. Hot Working Technology, 2017, 46(11):21-24 (in Chinese). [4] GAO X D, CHEN Y Q, YOU D Y, et al. Detection of micro gap weld joint by using magneto-optical imaging and Kalman filtering compensated with RBF neural network[J]. Mechanical Systems & Signal Processing, 2017, 84:570-583. [5] CUI W, DAI G, LONG F F, et al. Visualization method for the steel plate butt weld based on magnetic flux leakage testing[J]. Nondestructive Testing, 2013, 35(5):8-11 (in Chinese). [6] ZHEN R H, XIONG J B, ZHOU W. Micro-weld seam magneto-optical imaging regular research based on magnetic charge theory[J]. Electric Welding Machine, 2019, 49(7):84-88 (in Chinese). [7] TSUKADA K, MAJIMA Y, NAKAMURA Y, et al. Detection of inner cracks in thick steel plates using unsaturated ac magnetic flux leakage testing with a magnetic resistance gradiometer[J]. IEEE Transactions on Magnetics, 2017, 53(11): 2501305. [8] RICHERT H, SCHMIDT H, LINDNER S, et al. Dynamic magneto-optical imaging of domains in grain oriented electrical steel[J]. Steel Research, 2016, 87(2):232-240. doi: 10.1002/srin.201500013 [9] FANG J H. Analysis and control of ordinary welding crack in shipbuilding[J]. Ship Engineering, 2015, 37(s1):186-190(in Chin-ese). [10] WAN X Zh, ZHAO X Zh, LI M. Effects of cruciform butt welds on the static mechanical behavior of spliced circular steel tube[J]. Building Structure, 2011, 41(s1):843-846 (in Chinese). [11] GE J H, LI W, CHEN G M, et al. Analysis of signals for inclined crack detection through alternating current field measurement with a U-shaped probe[J]. Insight-Non-Destructive Testing and Condition Monitoring, 2017, 59(3):121-128. doi: 10.1784/insi.2017.59.3.121 [12] MA N J, GAO X D, ZHOU X H, et al. Analysis of magneto-optical imaging characteristics of weld defects under magnetic field excitation[J]. Laser Technology, 2018, 42(4):525-530(in Chinese). [13] GAO X D, LAN Ch Zh, CHEN Z Q, et al. Dynamic detection and recognition of welded defects based on magneto-optical imaging[J]. Optics and Precision Engineering, 2017, 25(5):1135-1141(in Ch-inese). doi: 10.3788/OPE.20172505.1135 [14] RAHMAN S, RAHMAN M M, HUSSAIN K, et al. Image enhancement in spatial domain: A comprehensive study[C]//International Conference on Computer & Information Technology. New York, USA: IEEE, 2014: 368-373. [15] MAINI R, AGGARWAL H. A comprehensive review of image enhancement techniques[J]. Computer Science, 2010, 2(3):8-13. [16] SUN Z G, WANG S J, ZHOU Y X, et al. Research on real time seam tracking technology based on image processing[J]. Manufacturing Technology & Machine Tool, 2019, 683(5):153-157(in Chinese). [17] CHENG H D, SHI X J. A simple and effective histogram equalization approach to image enhancement[J]. Digital Signal Processing, 2004, 14(2):158-170. [18] ZHENG W, SUN X Q, LI Zh. Image fusion based on Shearlet transform and region characteristics[J]. Laser Technology, 2015, 39(1):50-56(in Chinese). [19] GAO Y, WANG A M, WANG F H. Application of improved wavelet transform algorithm in image fusion[J]. Laser Technology, 2013, 37(5):690-695 (in Chinese). [20] CHEN W B, NIU B, GU H B. Fusion algorithm of infrared images and visible images based on weighting[J]. Aeronautical Computing Technique, 2016, 46(5):73-76(in Chinese).