微型机焊接缺欠自动检测系统的研究

A Study of Microcomputer-based Automatic Welding Defect Inspection

  • 摘要: 本文叙述在微机扩展的图象处理系统上,进行焊接缺欠自动检测的尝试.提出了“模块差分法”来加强模糊x光片图象的锐化效果;运用了加权Freeman码以利某些几何图形的边缘轮廓跟踪及特征抽取;设计了多级识别系统——树状分类器,使识别过程简化、有效.

     

    Abstract: In this paper the available approaches of the automatic inspection of the weldingdefects using a developed microcomputer image processing system are discussed.A "group-difference method" has been presented to obtain better result for theimage sharpening of the blurred welding defect x-ray picture.A useful concept of the "Weighted-Freeman's chain code" is used for simplifyingthe boundary tracking and feature extraction of some geometrical configurations.A multi-level recognition system-tree classifier is designed.It simplifys the process of recognition and makes it more effective.

     

/

返回文章
返回