A GRID-BASE LINE ANALYSIS FOR STREET OCCLUSION REMOVAL AND BUILDING FACADE TEXTURING
Yawen Liu, Zhen Guan · 2010
The building facade model is the most prominent streetscape feature for rendering virtual street models. This paper aims to develop a method to identify street obstacles which hinder photo-realistic texturing of building facade. In order to achieve this, the presented method focuses on two research components. Firstly, we develop an effective grid-based line analysis method to differentiate image elements including building facade from the ones containing street obstacles. The image is divided into a number of grids with a regular size. The dominant facade’s geometry plays a significant role as scene knowledge to filter street obstacle by analyzing the numbers of parallel lines and the density of parallel line per each grid. Secondly we present a method to generate an occlusion-free texturing of building facade based-on planar affine transformation and bilinear interpolation. We evaluate the effect of the presented method using real street-view images.