图象处理在装药疵病自动检验中应用的研究

• 作 者: 蒋建伟 李录荫
• 机 构: 北京理工大学力学工程系,北京理工大学力学工程系
• 出 处: 北京理工大学学报
• 发表日期: 1990年03期
• 关键词: 疵病 图象处理 装药 图象预处理 弹药/装药质量 边界轮廓 滤波去噪 跟踪方法 特征提取 自动识别

应用计算机数字图象处理技术研究从弹药装药的射线透视图象中自动检验装药疵病的方法。提出用装药区定位、减影,滤波去噪、三值化等图象预处理方法增强图象中的疵病信息;运用边界轮廓跟踪方法对疵病进行特征提取;设计了树分类器,对各类疵病进行自动识别和分类。实验结果表明,采用文中的方法可自动准确地从装药射线透视图象中提取各类疵病,得到疵病的大小、位置、数量、类型等定量检验结果。

Using computer digital image processing technology to study the method of automatically inspecting loading flaws from radiographic images of ammunition charge. This paper proposes to enhance the defect information in the image by using image preprocessing such as positioning, subtraction, filtering and de-noising of the loading area, and to extract the feature of the defect by using the boundary contour tracking method. A tree classifier is designed, Various types of flaws automatically identify and classification. The experimental results show that the method can automatically and accurately extract all kinds of flaws from the fluoroscopic radiographs, and obtain the quantitative test results of flaw size, location, quantity and type.