CAJ | 학술논문

本文针对X射线焊缝图像的复杂性,提出了准确快速的提取焊缝区域和缺陷的方法。由于检测到的X射线图像具有对比度不高、光照不均、缺陷边缘模糊、图像噪声多、存在较大的背景起伏等缺点,首先对含有缺陷的焊缝图像进行一系列的图像预处理;然后采用自适应阈值分割算法来提取焊缝缺陷区域,为缺陷特征测量做准备工作;最后,在图像本身所包含的几何特征、灰度特征、结构信息、颜色信息等特征中,对缺陷图像进行几何特征的测量,以便对缺陷类型进行分类。结果表明,该方法能较准确地提取射线图像的焊接缺陷,并能较准确的获取该缺陷的几何和代数特征,具有良好的适应性和实用性。
본문침대X사선한봉도상적복잡성,제출료준학쾌속적제취한봉구역화결함적방법。유우검측도적X사선도상구유대비도불고、광조불균、결함변연모호、도상조성다、존재교대적배경기복등결점,수선대함유결함적한봉도상진행일계렬적도상예처리;연후채용자괄응역치분할산법래제취한봉결함구역,위결함특정측량주준비공작;최후,재도상본신소포함적궤하특정、회도특정、결구신식、안색신식등특정중,대결함도상진행궤하특정적측량,이편대결함류형진행분류。결과표명,해방법능교준학지제취사선도상적한접결함,병능교준학적획취해결함적궤하화대수특정,구유량호적괄응성화실용성。
This article in view of the complexity of the X-ray weld image, and puts forward the method of accurate and fast extracting Weld area and defects. Because the X-ray images detected with low-rise contrast, the uneven illumination, blurry defect edge, more image noise, existing larger background ups and down and other drawbacks, first, we make a series of image preprocessing to the weld defect image containing defects;then adapting adaptive threshold segmentation algorithm to extract of weld defect area, which prepare for measuring defect feature work;Finally, the image itself contains the geometrical characteristics, the gray characteristics, structure information and color information, we select measuring defect image geometric features , so that to classify defect type. The results show that the method can accurately extract the images of ray welding defects and accurate for the defects of algebra and geometry characteristic, which has a good adaptability and practical utility.