A Research On The Key Technology For 3D Reconstruction Algorithm Based On The Feature Of Building Spatial Structure

Construction of Digital City is in line with the current status of the economic life of industrialization and information in parallel, which is the need of sustainable urban development and urban political and economic development. In urban planning, Digital City provides a new means of regulating urban planning, construction and management. As buildings are ubiquitous in the city, three dimensional modeling of buildings is the foundation of Digital City modeling. However, in the existing building modeling method, the wire frame model of the building cannot be generated automatically by the sparse shape points of the building space. Under this background, this paper presents a fast and simple algorithm for 3D reconstruction algorithm based on the feature of building spatial structure. The main research contents of this paper are as follows:(1) A new algorithm of building ground contour generation based on point by point is proposed. According to the disorder collection points, ground contour cannot be generated automatically. To solve this problem, the algorithm inserts shape points into the simple polygon which has been constructed by greedy algorithm in turn and automatic generates the building ground contour.(2) A new algorithm of building ground contour generation based on genetic algorithm is proposed. As traditional genetic algorithm is unable to generate a simple polygon directly, this paper proposes a new mutation operator and a local optimization operator in genetic algorithm, which significantly improves the accuracy of the traditional genetic algorithm to search for the optimal polygon.(3) Based on the idea of CSG model, a new algorithm of building wire frame model based on spatial sparse shape points is proposed.(4) A set of building simulation software is developed by using OpenGL which realizes the simulation and rendering results of the building wire frame model.