计算机视觉computer vision 11.docx

Chapter 11Shape from Photographs: A Multi-view StereoPipelineAbstract. Acquiring 3D shape from images is a classic problem in Computer Vision occupying researchers for at least 20 years. Only recently however have theseideas matured enough to provide highly accurate results. We present a complete algorithm to reconstruct 3D objects from images using the stereo correspondence cue.The technique can be described as a pipeline of four basic building blocks: cameracalibration, image segmentation, photo-consistency estimation from images, andsurface extraction from photo-consistency. In this Chapter we will put more emphasis on the latter two: namely how to extract geometric information from a set ofphotographs without explicit camera visibility, and how to combine different geometry estimates in an optimal way.11.1 IntroductionDigital modeling of 3D scenes is becoming increasingly popular and necessary fora wide range of applications such as cultural heritage preservation, online shoppingor computer games. Although active methods remain one of the most popular techniques of acquiring shape, the high cost of the equipment, complexity, anddifficulties to capture color are three big disadvantages. As opposed to active techniques, photograph-based techniques provide an efficient and easy way to acquireshape and color by simply capturing a sequence of photographs of the object.The goal of any shape-from-photographs algorithm can be described as “given a set of input photographs, how to estimate a 3D shape that would generate the samephotographs, assuming same material, viewpoints and lighting conditions”. This definition highlights the main difficulty of the problem: photographs are obtained as a result of complex interactions between the geometry of the scene, the materials of the scene, the lighting conditions and the viewpoints (see Figure 11.1).Hence recovering the geometry just from photographs is not only a challenging problem but also, in the general case, an ill-posed