Study On Detection Technology Of Crankshaft Based On Reverse Engineering

Reverse engineering is to meet today’s society rapid development of manufacturingand the rise of a kind of rapid manufacturing technology. But the whole processcontains a number of key technologies in reverse engineering, such as point cloud datacollection, data matching, and data error analysis, etc., these become the main problemsin study of reverse engineering technology. The crankshaft is a critical foundation forcomplex parts of cars, trains, ships and other large machinery products. In this paper,the crankshaft detection based on reverse engineering technology was studied.The maincontents are as follows:Firstly, crank model matching technology research.A matching method for pointcloud data and original design data model was presented based on the crankshaft feature.The method of the matching of the crankshaft is divided into two steps:The initialmatching method based on feature points of crankshaft and exact match based onimproved ICP algorithm.Select the calibration three feature points of the crankshaft anduse the relationship of these points to strike a transformation matrix, Initial matching ofthe two models will finished,then build a KD tree data structure to accelerate the nearestpoint,and finally to achieve an exact match.Secondly, Second, the curvature error of a crankshaft point cloud data analysismethods. The method by calculating curvature of point cloud data and the originaldesign data, analysis and comparison the curvature of the corresponding points, analysiscorresponds to the point of curvature error,Distance error can be directly clear that the distance between two points, thecurvature error can be directly shows that the shape error of the model, the combinationof two methods can get better detection of the crankshaft.Finally, for the scan blindness crankshaft clouds data in reverse engineeringacquisition process, this paper presents the method of detecting the scan blindnessthrough two-dimensional pictures.The method finishs the detection of the blind throughthe relation between object in the world coordinate system, the image coordinate systemand camera coordinate system to establish the mathematical model between the pictureand the3D entity. Moreover, this method also can detecte the object directly, mainlysome distance, depth and position of a feature detection can be achieved.The actual comparison of the crankshaft test results verify the correctness and validity of the conclusions of this study.