Reconstruction Of Three-dimensional Model Based On Point Cloud Data

Three-dimensional model is the simulation of the real world. People could accomplish some significant works such as urban planning, industrial design and heritage restoration by means of three-dimensional model. In particular, the advent of three-dimensional laser scanning technology makes the reconstruction of three-dimensional model really much more easier to come true. Using the method which is called real replication technology could get vast amounts of three-dimensional point cloud data, then we could complete the transformation from the raw point cloud data to real three-dimensional model by a series of post-processing. Due to this three-dimensional model reconstruction technique based point cloud data have many advantages that those traditional measurement methods could not match, some problems exist in the present as well. So a large number of scholars study in this field, and this technique becomes one of the focus of surveying and mapping disciplines in the current.The test studies some current problems that exist in the three-dimensional model reconstruction technique based point cloud data, and puts forward the corresponding solution to these questions.Firstly, In view of the commonly used smoothing algorithm may change the coordinate value of the points, but other filtering algorithm without changing the coordinate values of points often need more than coordinate information such as reflection intensity. A new filtering algorithm based on distance determination is put forward in the paper. The new algorithm introduces statistics theory, and only need to provide the point coordinate values. Validating the result of new algorithm by developing homologous program, then compared with the photo of original point cloud data, It is not difficult to find that new algorithm has good filtering effect, and could realize the point cloud data filtering.Secondly, Regional gravity compression algorithm is earlier raised than other compression algorithm in the procedure of the point cloud data compression, and it is commonly used in post-processing. For the phenomenon that it may lost much detailed features of model when this algorithm be used, the improved regional center compression algorithm is brought forward in this paper. The improved algorithm divides the procedure of compressing point cloud data into two steps and identifies the value by a threshold in each of the steps. it not only retains the gravity point of the grid body but also keeps back the apogee in line with a certain threshold range to complete the initial compression, and then the compression of the point cloud data is finished. Validating the result of the improved algorithm by developing homologous program, then compared with the result computed by regional center compression algorithm. It is not difficult to find that the result computed by improved regional center compression algorithm is better than another in the same conditions, and it would save more detailed features when the former algorithm used. the result above prove that the improved regional gravity compression algorithm is reliable.Finally according to the present modeling platform is imperfect and incompatible, and other issues exist in as well, It is the multi-platform combined modeling techniques that summarized on the basis of analysing the advantages and disadvantages of the modeling platform like Geomagic Studio, Cyclone, PCL and auxiliary modeling software like Photoshop, 3DMax, SkyLine in this paper. This method takes best use of the advantages of various software and integrates all of these advantages in modeling process, making the result of model reconstruction better than others’. In addition, The article gives some ideas about data format conversion between the various platforms by reconstructing three-dimensional model for parts of the building in Lin Tong campus of Xi An University of Science and Technology.