3D Laser Colored Laser Scanning And Range Detecting System Design And 3D Point Cloud Data Model Reconstruction

With the rapid development of 3D scanning and ranging technology,3D point cloud is more and more widely applied in the field of industrial detection, autonomous navigation, reverse engineering, etc. In the process of digitalizing the real world,3D point cloud data can record geometric properties and location message of the object's surface accurately while the 2D image records the color and texture. The close integration of the two will produce a new kind of digital media, namely 3D colored point cloud data.3D colored point cloud data is the upgrade of 3D point cloud data as it possesses not only the geometry and location information of the object's surface but the color and texture of that, with which it can illustrate the real world more precisely.In order to obtain 3D colored point cloud data, this thesis has designed a new 3D colored laser scanning and range detecting system, mainly composed of 2D laser scanning rangefinder, high-speed industrial camera, electric turntable and host computer.2D laser scanning rangefinder and high-speed industrial camera are installed onto the electric turntable. When launched, it will rotate at a constant speed for the two to carry out the omnidirectional collection of 3D point cloud data and 2D image data. The host computer gets 3D point cloud data and 2D image data through Ethernet communication, and conduct precise integration to obtain omnidirectional 3D colored point cloud data. In this thesis, VC++and OpenGL are employed to write a software to collect and integrate 3D colored point cloud data, attempting to integrate and display 3D colored point cloud data.3D point cloud data processing technology has played a key role as the basis of a variety of applications. Model reconstruction is a very important technology in 3D point cloud data processing. Considering the non-uniform of 3D point cloud data collected by the scanning and range detecting system, the author puts forward a kind of triangular mesh modeling method to the equilibrium. This method firstly uses KD-tree to search the neighborhood of every point the in 3D point cloud, and employs the covariance matrix to calculate the normal vector and the tangent plane, then projects the neighborhood of a given point to its projection plane, and through the fan zoning and anisotropy equilibrium selection, further optimize the cutting plane projection of the neighborhood. Finally, adopt the method of two-dimensional Delaunay to carry out triangular mesh modeling on the optimized projection neighborhood, and reversely project the built 2D triangular mesh model to the 3D neighborhood space, in order to realize the 3D triangular mesh model reconstruction of point cloud. Compared with the reconstruction method of existing model, this method use fan-shaped zoning and anisotropy equilibrium selection to better choose the neighboring points of a given point in all directions, making the neighborhood of it distribute homogeneously in all directions and realizing accurate modeling of the heterogeneous 3D point cloud.