Contour Recognition Research For Metallic Parts Deposition Shaping Based On Laser Generator

Arc Deposition Manufacturing(ADM), not only shares the inherent advantage ofadditive manufacturing, but also has the ability of forming component with excellentmechanical performance, which get wide attention from researchers. While the progress ofADM is a complex physical phenomena, It is hard to insure the performance of depositionlayer in the process of metallic parts additive manufacturing ifdeposition processinformation can’t be collected on line and it will seriously affect the precision andperformance of the final deposition part. In order to get deposition part with high qualityand high precision, it is necessary to monitor the deposition process on line and adjust thetreading track, parameters of the deposition current voltage and arc length etc.Firstly, To meet the requirement of the deposition layer morphology detection duringmetallic parts additive manufacturing a kind of deposition layer morphology detectionsystem using active light source laser is proposed, Installation way with laser generatordip sweeping deposition layer is adopted considering the size of assembly space. Abenchmark slice is designed and installed on the detection device to keep the datum line inthe field of view. The height information can be worked out based on the offset betweenthe position of the laser rays on deposition layer and on slices. The total height to thebaseplate of points on deposition layer can be figured out for two measurement waysdescribed above.Then, The curves of height of multichannel transverse deposition layer is worked outbased on the isometric sex of consecutive images, a kind of three-dimensional shapereconstruction algorithm is exploited using the calculated height information for singlelengthways deposition layer, the intrinsic parameters of CCD is calibrated based on thematlab tool, In addative, The actual physical size dx、dy of single pixel in the workpsplaceis calibrated and the intersection angle between lasing area and optical axis of CCD iscalculated; After median filtering, twice morphological opening, image segmentationbased on maximum entropy, canny edge extraction, a binary image full of the informationof laser rays can be obtained, algorithms with strong adaptability extracting the laser rayson the benchmark and deposition layer is developmented which insures the reliability offollowing height calculation and three-dimensional shape reconstruction.Finally, experiments are implemented on a single lengthways deposition layer underthe affect of magnetic field and a rotational parts with thick wall. The results show that thereconstructed three-dimensional shape for single lengthways deposition layer is close tothe real deposition layer and the height error is in0.55mm on the10th layer.To therotational parts, The curves of height can accuratelyreflect the fluctuation of the surfaceand the height error is in0.7mm.Both can privode valuable data for trajectory optimizationof the next layer. It is concluded that the detecting system can work well on the depositionlayer morphology detection and has practical engineering value.