毋立芳, 高源, 郭小华, 赵立东, 张子明. 模型引导与视觉分析相结合的3D打印产品质量检测方法[J]. 北京工业大学学报, 2017, 43(4): 526-532. DOI: 10.11936/bjutxb2016070040
    引用本文: 毋立芳, 高源, 郭小华, 赵立东, 张子明. 模型引导与视觉分析相结合的3D打印产品质量检测方法[J]. 北京工业大学学报, 2017, 43(4): 526-532. DOI: 10.11936/bjutxb2016070040
    WU Lifang, GAO Yuan, GUO Xiaohua, ZHAO Lidong, ZHANG Ziming. Quality Control Approach of 3D Printing Products Combining 3D Model and Computer Vision[J]. Journal of Beijing University of Technology, 2017, 43(4): 526-532. DOI: 10.11936/bjutxb2016070040
    Citation: WU Lifang, GAO Yuan, GUO Xiaohua, ZHAO Lidong, ZHANG Ziming. Quality Control Approach of 3D Printing Products Combining 3D Model and Computer Vision[J]. Journal of Beijing University of Technology, 2017, 43(4): 526-532. DOI: 10.11936/bjutxb2016070040

    模型引导与视觉分析相结合的3D打印产品质量检测方法

    Quality Control Approach of 3D Printing Products Combining 3D Model and Computer Vision

    • 摘要: 针对3D打印产品个性化程度高的特点,普通的产品质量检测方法并不适用. 对3D打印额面肿瘤导板导向孔,为取得更高精准的检测,提出了一种模型引导、机械系统配合、视觉系统监控相结合的质量检测方法. 整个方法框架包括机械系统、控制模块、视觉模块,在检测算法的统一调度下实现3D打印额面肿瘤导板导向孔精度检测. 首先,利用导板和相应三维模型上对应的各3个不共线点对齐导板和模型. 然后,根据三维模型上的导向孔位置和方向,控制机械系统引导模型导向孔移动到视觉模块的中心,通过分析导向孔的圆度、长宽比、直径、孔心间距等特征参数,判断导向孔是否合格. 结果表明:该算法和系统,相比于人工质量检测,缩短了检测周期,提高了检测精度. 提出的算法框架可用于其他3D打印产品与模型的形状一致性检测.

       

      Abstract: The general sampling inspection solution is not suitable for 3D printing products because of the personality of such products. In this paper, a framework for 3D printed maxillofacial tumor treatment guide inspection was proposed to achieve a more accurate detection, which includes mechanic system, control part and video surveillance part. The proposed method was used to inspect the location and direction of the hole under three parts. First, the 3D model and the printed product were aligned in the unified coordinates. Then, the product was moved to the suitable view based on the location and direction of the hole on the 3D model so that the hole would be in the center of the image. Finally, the hole was detected and shape parameters such as circularity, aspect ratio, diameter and the distance between the neighboring holes were extracted. Based on these parameters, we can decide if the hole is qualified. The experiments demonstrate that the proposed method raises the inspection precision and speed, in comparison to manual inspection. Furthermore, the proposed framework can be extended to shape consistency inspection of other 3D printing products.

       

    /

    返回文章
    返回