摘要
提出利用非量测CCD摄像机进行钣金件高精度3维重建与制造误差检测。介绍直线段最小二乘模板匹配的基本原理;提出利用像面上的点、直线信息进行混合光束法平差,对工业钣金件的线框模型进行高精度3维重建;采用1维模板匹配技术直接获取复杂形状的物方参数;重建结果可以用于检测零件的制造误差。所开发的误差检测系统具有硬件成本低、自动化程度高等特点,实际图像数据的多次实验均取得很好的检测结果。
A new approach of three-dimensional reconstruction and inspection of industrial sheetmetal parts with non-metric CCD camera is proposed. Principle of line segment least squares template matching to extract precise points and lines from imagery is discussed. Wire-frame model of the sheetmetal part can be accurately reconstructed with hybrid point-line photogrammetry. One-dimensional template matching and direct object space solution is used to reconstruct complex shapes such as circles and connected arcs and lines. Producing imprecision can be inspected automatically or interactively by the results of reconstruction. The proposed inspection technique has the advantages of low cost of hardware and can run automatically and fastly. Inspection results of several parts are very satisfying.
出处
《测绘学报》
EI
CSCD
北大核心
2004年第2期132-137,共6页
Acta Geodaetica et Cartographica Sinica
基金
国家自然科学基金资助项目(40301041)
关键词
72序列图像
非量测摄像机
3维重建
误差检测
最小二乘模板匹配
钣金件
误差检测
image sequence
non-metric camera
3D reconstruction
imprecision inspection
least squares template matching
hybrid point-line photogrammetry
