Photogrammetric method of virtual model generation for complex 3D objects

For realistic virtual model generation two main problems have to be solved: to achieve the required geometric accuracy in object description and to obtain the highest reachable texture similarity. Recently digital close-range photogrammetry offers precise and flexible means for the first problem solution accurate geometric 3D reconstruction. Besides images acquired for 3D reconstruction can be used for adequate photometric modeling generation of realistic object texture. The close-range photogrammetry technique for accurate geometric and photometric modeling is well developed for 2.5D (depth) object models. But for complex 3D objects which have no single valued plane projection there are some problems with surface reconstruction and texture generation. In this case uniform surface can not be reconstructed by common Delaune triangulation procedure for all points of objects. So the orthophoto can not be generated because of surface singularity. The paper presents the approaches for these problems solution. It describes the method for reconstruction of the uniform surface of complex 3D objects having object space coordinates cloud and the method for accurate texture generation from a set of object images obtained from various view points. The results of objects reconstruction and texturing are given along with the description of GosNIIAS close-range photogrammetric system description.