The Research On Characteristics Filling Method Of The Scanning Point Cloud About Building's Edges And Corners

Three-dimensional laser scanning technology obtains a wide range of applications at this stage for its speed,high precision,real time and non-contact,etc.The development of laser scanning technology provides a new technical means for spatial information acquisition and urban construction,making people convert from the traditional manual single-point data acquisition to automatic continuous access the bulk data,and improving the measurement accuracy and speed.Because in the process of 3D Laser Scanning,we conduct discrete sampling for scanning the object in terms of certain angle resolution to cause the angular features of object surface point cloud fuzzily.Meanwhile,the information of buildings'surface and ridge needs to be identified quickly when we use laser point cloud data for building model and vector information extraction.Based on this,this paper starts from the terrestrial laser scanning point cloud data and carries out a deep study on the fast extraction method about the buildings'surface from point cloud data to realize filling the missing feature points fleetly and accurately from point cloud data.The main research contents are as follows:(1)Propose a fast extraction method research about the buildings'surface and ridge based on shared Nearest Neighbor algorithm.The main steps are as follows:first,calculate the normal vector for each data point of point cloud and use Grid-based shared Nearest Neighbor algorithm to classify point cloud to ensure point cloud on the same surface of the buildings;then automatically distinguish intersecting plane and extract the ridge of buildings,and fill the ridge on the missing feature points.The result shows that the degree of automation of the method is high and the extracting of the buildings'surface and ridge is fast and accurate,the extraction results can be applied to automatic modeling of three-dimensional buildings and draw the map.(2)Propose a high-precision splicing method of urban point cloud based on the buildings'ridge,utilizing the extracted buildings'ridge.The method firstly utilizes the characteristic that the corner angle of urban buildings is on plumb line,automatically extracts public straight edge of the same buildings from the point cloud of adjacent scanning stations,confirms attitude parameter of two stations point cloud?and?,corrects the point cloud horizontally for all station tilted point cloud and ascertains the conversion parameters of plane of the two stations'same vertical lineX_S and _SY;after corrected,we use the buildings'straight ridgeline of public edge between adjacent point cloud to compute the horizontal swing parameters?and the conversion parameters of heightZ_S.Experiments show that the method adopted in this paper has stronger operability and higher working efficiency,the splicing precision can achieve 2mm,an order of magnitude higher than the splicing precision of ICP algorithm.(3)Discuss the effect of the fill in ridge points on 3D model of the object,according the point cloud which the characteristics have filled to build 3D model on the geometry shape and size,similar to the real objects;elaborate the application of feature extraction in data compression and architectural decoration industry to lay the foundation for the application of 3D laser scanner in architectural decoration industry.