RI: Small: Modeling Cities by Integrating 3D and 2D Data

点击翻译按钮获取中文摘要

This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5).The problem of automated 3D reconstruction and modeling of urban environments is of great interest and importance in the fields of computer vision and graphics. This is due to the fact that accurate 3D city models are paramount in the further development of a variety of fields. This project acquires vast amounts of data via the latest generation in laser scanning technology and high-resolution cameras in order to achieve the goal of 3D city modeling. That data is registered in a common frame of reference based on current techniques that are improved in order to facilitate voluminous point clouds. One of the major challenges is the complexity of the acquired dataset that needs to be simplified for efficient rendering and higher-level recognition algorithms. Current simplified methods produce reduced sets of triangles that lack higher-order labels. This project segments and classifies the data into coarse and fine urban elements (such as facades, windows, vegetation, etc.), discovers symmetries among the elements, and fills missing data. It also integrates image-based and laser-based models in order to enhance the acquired geometry, and to produce a seamless visualization result. The above will also aid recognition processes, alternative visualizations and truly photorealistic representation of the 3D scene. This work is disseminated through collaborations with the geography and film-and-media departments of Hunter, as well as other agencies (such as museums). It has significant impact for urban planning, architecture, and archeology applications, and in systems such as street map visualization, as well as in film and construction industries.