Research On Feature And Constraint Based Visual Reverse Technology For Two Dimensional Parts

As an important part of modern manufacturing technology, reverse engineering hasbecome an important technical means to extract product design parameters and designintent, and has become an important way to new product development, which has abroad application prospect. Model reconstruction theory, as the key technology ofreverse engineering, is always the key point of research. This paper carries on relatedwork in study and implementation of relevant reconstruction theory based on contourfeatures and constraints.The image of the target parts is taken using the image acquisition platform in thelaboratory at first. When get the image of parts, point cloud data of parts’ contours canbe extracted after filtering, denoising, binarization, single edge pixels extraction. Then,focus on the segmentation method of point cloud data. By comparing the advantagesand disadvantages of curvature based region segmentation method and non-curvaturebased region segmentation method, an algorithm taking straight line and arc as theprimitive element to approximate contours is advance to segment cloud data. The resultthat straight line and arc are the basic primitive is got. By removing feature pointswhere the constraint relation between adjacent primitive is a tangent, the segmentationresult that straight line, arc and b-spline are the basic primitive is got further. Based onthe recognition criterion to the three kinds of elements, the the optimum curve type ofevery data segment is got. And the parameters of each segement is got at the same time.Based on the constraint relationship criteria, the cinstraint relaitionship between thecurves is recognized later. Finally, the contours of part is reconstructed through theoptimum mathematical model based on feature and constraints.Corresponding algorithms in this article are realized by VC++. Feature points ofpart contours are extracted, on the premise of meeting the design intent. The contours ofpart is reconstructed based on the recognition criterion, the constraint relationshiprecognition criteria and the optimum mathematical model considering feature andconstraints. The feasibility and correctness of the reverse solving method is verified byreconstruct actual parts.