Laser Data Fast Registration Drawn

The 3D Laser Scanner has broad application in heritage preservation, urban planning, digital cities and the mine surveying due to its convenience. These data that obtained by the Laser Scanner is called Depth Data, because they contain the distance information between from the scanner to screen. At the same time, using the camera which provided by Laser Scanner or professional photographic equipment can also obtain the two-dimensional texture information of object, which can reconstruct the real world.However, the high complexity of the real world results in many difficulties and challenges. One of them is the registration of the range images. In order to obtain a whole 3D model of a screen, we must align scans form different viewpoint. Although lots of has been done on image registration of small objects. Large data sets and the great number of scans make these methods difficult to apply to large real scenes. In addition, the screen render has becoming more and more complex, and the render rate of which people expected has becoming higher. Although the traditional rendering can produced highly quality images, but it faced great challenge in the aspect of real-time rendering due to its own limitation.This article has first discuss the weak of the traditional ICP point to the actuation of image registration ,and then Base the initial registration ,this improved ICP algorithm leads to a fewer time and more accurate effect after K-D tree on finding the closed points. an idea about the program design is also presented. In the aspect of Scene Rendering, First we introducing the algorithm of Ray Casting, Splatting, Shear-Warp and so on. Then we have done volume rendering using the algorithm of Marching Cubes and the Ray Casting based on the Visualization Tool Kit point to the data contained 3D point data and image data, and achieved good resultsAt last, the article also researched data acquisition, based on Laser Scanner, the pre-processing of point cloud and volume rendering based on VTK and so on.