| Popis: |
In this paper, we propose an improved algorithm for 3D scene reconstruction from images obtained from two or more cameras. The principal objective of the proposed algorithm is to reduce number of feature points used for 3D scene reconstruction by using a triangle mesh and recursive sub-mesh definition. This kind of optimization can reduce computational complexity and the algorithm thereafter might be applied in various real-time applications. Another benefit of the solution is reducing the need of computing power, hence a less powerful and more energy efficient computers can be placed e.g. in vehicles. At the beginning of this paper we introduce several issues of 3D reconstruction algorithms, the middle part is focused on the description of the proposed algorithm and in the conclusion we outline a few potential algorithm applications. |
