3-D Reconstruction Of Buildings Based On LIDAR Data And Aerial Image

With the deeply application of 3-D GIS and virtual realism, conventional 3-D modeling methods such as engineering survey and photogrammetry were not seasoned with current require because of its'scarity in work efficiency and data precision. Lidar (Light Detect And Ranging) techonology is one kind of high technology integration system,which can explor objects and measure distance by the way of laser scanning.With this system wen can obtain urban data surface model easily. Another choiced way for urban house reconstruction was occurred when the technology was succeeded. Conventional aerial picture, which was obtained through aerial photoing, can get the information of objects according to parameter of camara and geometry ralation of the object and camera.This paper demonstrates and analyzes Lidar data and aerial image. The first step is the region threshold segmentation of Lidar data and extraction of the edges of apparent buildings, wiping off groud area. At the same time, we must remove the wall dots that should not exist. We all should remove the dots that does ont belong to house area, because what we are interested in was the house area. Secondly we should process the aerial image. Using canny edge dector to do with aerial image and then we get the two-value picture. After then apply HOUGH transform to the picture to acqure the line information. In fact, it is very hard to avoid error during the course of aerial photo working, the offering data is not according with the realism data. It is necessary to compute the image parameter by the method of backward intersection. The project effect of the Lidar dots to the picture will be fine after parameter rectifing. The parameter computing is just one kind of data matching course in so many woeds. On the base of matching project the segmentated LIDAR data to the aerial image. Here the LIDAR data just lie within the contour of the house. Then search the line that extracted which was nearst to every contour dot. Lastly, ensure the topology relation of the searched lines and construct 2-D model. The 3-D constructing is contrary according to the last step. Get the dot that is the nearst to each contour dot and endow the contour dot with the lidar' height value. At last, each vector special coordinate value can be computed according to the co—linear equation.In one word, ideas of this paper were to apply the height value of LIDAR data and the contour information of the house in the aerial. Both sides supplement each other This way is superiory according to conventional photogrammetry methods.