Research On Automatic Detection Methods For Line Heating Complex Deformation Surface

Line Heating technology has been widely used in the manufacturing of molding complex curved plate surface for a long time. The automatic test and discrimination of molding curved surface is the key to the automatic forming and processing of hull curved plates. At present, the detection during the production of a shipyard is completed with the model or sample box chosen by skilled workers based on their rich experience. The manual way of Line heating detecting will bring not only low efficiency but also big errors, which can’t meet the requirements of the modern shipbuilding industry on automatic processing of board building. And this has been the bottleneck to restrict the cycle and quality of ship production. The automatic discrimination for molding curved plate surface of hull boards is not only an urgent issue, but also an important factor to realize the automatic processing of Line Heating.This paper proposes a automatic discrimination method for molding curved plate surface of hull boards based on error estimation of contour profile. The main tasks of this paper as follows:(1) Using laser measurement to get the3D point cloud data of the hull plate surface formed by Line Heating, and then those point cloud data follows the processing of denoising, smoothing and reducing. The denoising of the point cloud data can effectively inhibit the influence of noise data points on curved surface reconstruction and matching precision; while the point cloud streamline technology greatly reduces the amount of data, which is preparing for the reconstruction and testing of the curved surface.(2) This part first introduces the basic idea and principle of isosurface extracting in Marching Cubes algorithm, and then proposes the improved method after analyzing the cause of low efficiency in MC algorithm. At last using the improved MC algorithm to reconstruct forming curved surface of the hull plate after laser scanning.(3) This part does some research on the error analysis theory of curved surface’s profile and the basic principles of ICP (iterative nearest point) algorithm firstly, and proposes a neighborhood-relationship-based search algorithm to search corresponding points, using quaternions method to solve out the optimum matrix in ICP algorithm. In consideration of the existing defects of the basic ICP algorithm, using two footwork matching ways to match point cloud data, and then proves the match algorithm with the pommel horse curved surface.(4) Summarizes the VTK framework, the3D rendering principle and basic graphics rendering steps, develops the free curved surface detection system under the Visual C++6.0environments using object-oriented technology. And then realizes the reconstruction and the matching process of curved surface formed in Line Heating as mentioned in the first two chapters.To sum up, with the laser measurement technology as the foundation, taking the detecting of curved surface of hull plate as the starting point, provides a method to reduce the error of3D curved surface in Line Heating through error analysis and compensation for the applying curved surface, and then provides a theory basis for solving these problems occurred in real practice. The experiment shows that it is feasible to realize precisely detecting through the surface reconstruction and surface matching technology, The study has certain practical value.