Key Technologies Research On Digital On-line Inspection System For Complex Surfaces

With the development of CAD/CAM/CAI integrated technology, using On-line inspection for the free-form surface quality inspection process is a trend. On-line inspection can save the expense of facilities and measuring time cost, avoid the fixture change for workpiece, Besides, the integration of inspection,design and machining process is satisfied with the aim of modern manufacturing. Now the engineering applications and basic theories researches about on-line inspection has performed from various aspects, In this paper, the discussion and research on the key the technologies for the On-line inspection system is carried out, We mainly consider the four key question:CAD model's visualization, the representation of free-form surfaces, measuring point sampling methods, Algorithms for accurate surfaces matching.At first, the theory of On-line inspection system is analyzed. According to the analysis for the modules of the system to determinate the technical lines and propose the scheme for system.Secondly, based on the result of theory analysis, many key technologies for system is deeply discussed, including:1. The visualization of CAD models and Free-form surface modeling. An interpretation of STL CAD models is presented and the models visualization is realized with no information lost. A good human-computer visualization environment is established by the powerful function of OpenGL. Using the bi-cubic NURBS method to represent the surface for inspection, the surface's C2 continuation and the smoothness is ensured, then the surface is represented accurately.2. The scheme for measurement point sampling. The problem of single sampling method's low adaptability is proposed by synthetically analyze these existed methods, and a new sampling strategy is presented by considering the shape features. Based on the differential geometry theory to analyze the shape features, including curvature feature and surface patch feature, and then using the features to find suitable sampling method. The strategy can utilize the advantages of these methods. Besides, the path generation is discussed, proposing that if plan the path basing on TSP theory should take into account the time cost of probe head re-orientations, but it's hard to determine, so a practical method which performed in unidirectional or bi-directional is proposed, where the former is generally better suited to closed surfaces and the latter is more suited to open surfaces. The sampling points' normal vector is calculated by Deboor algorithm and the collision free path problem is discussed. 3. Surfaces best matching. The problem of best matching for two surfaces is resolved by ICP algorithm. About the close point for dimensional point to a surface, a algebraic derivation process based on Newton iterate method is presented, besides, a high efficient non-iterative method for solving the objective function is applied base on singular value decomposition.At last, an inspection software based on above theories is developed, and then the whole system platform is established. The inspection experiment for parts of typical complex surface is carried out to verify the system.