This paper presents an interactive modeling system that constructs 3D models from a collection of panoramic image mosaics. A panoramic mosaic consists of a set of images taken around the same viewpoint, and a transformation matrix associated with each input image. Our system first recovers the camera pose for each mosaic from known line directions and points, and then constructs the 3D model using all available geometrical constraints. We partition constraints into soft and hard linear constraints so that the modeling process can be formulated as a linearlyconstrained least-squares problem, which can be solved efficiently using QR factorization. The results of extracting wire frame and texture-mapped 3D models from single and multiple panoramas are presented.