Based On Monocular Visual Aids Of The Three Coordinate Measuring Machine Measurement Trace Planning

The measurement of curved surface with high efficiency and precision is always the major research emphasis in industrial circle of automobile, mould, and component and parts detection, etc. The thesis regards unknown CAD model surface of CMM measurement as the object of research, a reconstruction model of the object to be measured is established by SFS method based on monocular vision, which can be used to guide measuring path planning on CMM. The thesis contains the main contents as follows:(1)A3D CAD model of the object to be measured of is established by SFS method, which size is repaired and reference is merged according to the characteristic of the model. Based on similarity theory and equal proportion relationship,2D size of SFS reconstruction model is exactly reconstructed; through reference coordinate transformation under different situations, achieving reference fusion between SFS reconstruction model coordinate system and machine coordinate system.(2)To solve the irrational problem of distribution of the measured points in the traditional surface sampling method, this thesis proposes a sampling method based on curvature chord height method. The first considers the factor of curvature changes in the curve, the chord height function is built, every measuring point changes of chord height value can be obtained; based on the principle by which the data simplification of chord height method, the reserved sample point is determined to achieve the reasonable distribution of measuring point based on the curvature of surface; according to the ratio between true point and predicted point, the true value of measuring point is predicted, the true value of measuring point is accurately predicted and sampled by means of measured value feedbacks.(3)A path planning method based on layering principle and hill climbing genetic algorithm is proposed, which can enhance measuring efficiency and at the same time avoid a collision of probe. Geometry division is carried out with measured surface according layering theory, then hill climbing genetic algorithm is used for optimal computation of measuring paths from measuring point in different areas, and the path in different areas is integrated to obtain the full path planning.(4)A prototype system of curved surface measuring path planning is developed. The measuring path planning is employed for solving the measuring about mouse, fan blade, structural component and planning the measuring path, which also compared with traditional measuring method on precision and efficiency, tested the feasibility and high efficiency of the method presented in this thesis.