基于灰色理论的X射线气缸套缺陷自动分类技术研究

发布时间：2018-12-31 08:31

【摘要】：气缸套作为内燃机的关键重要部件,国内外厂家与用户都对其生产质量提出了严格的要求。目前,针对气缸套缺陷检测的常用方法主要是涡流检测和X射线检测。然而,由于肌肤效应,涡流法只能检测到气缸套表层或近表层缺陷,对气缸套的深层缺陷难以检测,经涡流检测后的气缸套有可能会把隐患带入到后续的内燃机装配生产中。当前的X射线检测方法大多依赖于人眼识别,面对当前大批量生产的气缸套产品,这种人眼判读法不仅检测速度慢,人工成本高,而且易受主观因素的影响,可靠性不高,不利于后期对缺陷成因的分析与铸造技术与工艺的提高。基于上述背景,本文提出了基于灰色理论的X射线气缸套自动检测技术。针对缺陷特征难以提取的问题,本文首先利用简化B型关联度检测出气缸套的边缘和缺陷小块,然后利用形态学重建的边界填充方法对气缸套边缘进行填充,提取出缺陷小块,最后利用形态学的膨胀腐蚀对缺陷小块进行预处理,提取出气缸套中的缺陷。缺陷特征提取的难点是如何提取出缺陷小块,本论文经过分析缺陷区域属于高频分量,计算其中心像素和它周围8邻域像素差值的均值构造新的子图像矩阵,进而增强高频分量,然后采用矩阵均值化算子和简化B型关联度相结合的方法检测气缸套的边缘和缺陷小块,准确提取出气缸套的缺陷。产品的缺陷检测一方面用以保证产品的质量,另一方面对废品的缺陷分类识别则有助于找出缺陷成因,提高气缸套的铸造生产工艺,提高产品合格率。本文利用递进模糊识别方法对缺陷进行分类,主要根据缺陷的特征属性,判别气孔、缩松、缩孔和裂纹的隶属度阈值并进行对比,进而判断缺陷所属类别。实验结果表明,本文提出的方法实现了气缸套缺陷的X射线自动检测识别,可以快速准确提取出气缸套的缺陷,并实现了缺陷类别的快速分类。本论文相关技术的研究,有助于我国X射线自动识别技术的发展。
[Abstract]:As a key component of internal combustion engine, cylinder liner has been demanded by domestic and foreign manufacturers and customers. At present, eddy current testing and X-ray detection are the main methods to detect cylinder liner defects. However, due to the skin effect, the eddy current method can only detect the surface or near surface defects of cylinder sleeve, and it is difficult to detect the deep defect of cylinder sleeve. The cylinder liner after eddy current detection may bring hidden trouble into the subsequent internal combustion engine assembly production. Most of the current X-ray inspection methods depend on human eye recognition. In the face of the current mass production of cylinder sleeve products, this human eye interpretation method is not only slow in detection speed, high in labor cost, but also vulnerable to subjective factors, so its reliability is not high. It is unfavorable to the analysis of the cause of defect and the improvement of casting technology and technology. Based on the above background, this paper presents an automatic detection technique of X-ray cylinder liner based on grey theory. In order to solve the problem that the defect feature is difficult to extract, this paper firstly detects the edge of the cylinder sleeve and the defect piece by using the simplified B-type correlation degree, and then uses the boundary filling method of morphological reconstruction to fill the edge of the cylinder sleeve, and extracts the defect block. Finally, the defects in cylinder liner were extracted by morphological expansion corrosion pretreatment. The difficulty of defect feature extraction is how to extract the defect block. In this paper, the defect region is analyzed as a high-frequency component, and the mean value of the difference between its central pixel and its surrounding 8-neighborhood pixel is calculated to construct a new sub-image matrix. Then the high frequency component is enhanced, and then the edge and defect of cylinder sleeve are detected by using matrix mean operator and simplified B-type correlation method, and the defects of cylinder sleeve are extracted accurately. On the one hand, the defect detection is used to ensure the quality of the product, on the other hand, the classification and identification of the defects of the waste products can help to find out the causes of the defects, improve the casting process of the cylinder sleeve, and improve the qualified rate of the products. In this paper, the progressive fuzzy recognition method is used to classify defects. According to the characteristic attributes of defects, the threshold of membership degree of pores, shrinkage holes, shrinkage holes and cracks is judged and compared, and then the categories of defects are judged. The experimental results show that the method proposed in this paper realizes the automatic X-ray detection and identification of cylinder liner defects, and can quickly and accurately extract the cylinder liner defects and realize the fast classification of defects. The research of this paper is helpful to the development of X-ray automatic recognition technology in China.
【学位授予单位】：中北大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;N941.5;TK403

【相似文献】