一种基于深度学习的户型图矢量化方法

A Vectorization Method of Floorplan Images Based on Deep Learning

提取房屋结构是重建三维房屋的基础工作,主要被应用在虚拟现实(virtual reality,VR)看房、在线家装设计等新兴领域。针对户型图风格多样以及传统图像处理方法自动化程度低、普适性差的问题,提出一种结合深度学习和图像处理的方法,先通过深度学习方法获取墙、门和窗户在图像中的位置信息,再通过图像处理方法获得矢量化结果。在二维户型图上进行实验,结果表明了所提户型图矢量化方法的有效性和适用性。

Extracting the structure of a house is the basic work of reconstructing a 3D house.It is mainly used in the emerging fields such as virtual reality(VR)house viewing and online house decoration design.Aiming at the problems of various styles of floorplan images and the low automation and poor generality of the traditional image processing methods,we propose a method that combines deep learning and image processing.First,the position information of the walls,doors,and windows in the images are obtained by deep learning method,and then the vectorization results are ob⁃tained by image processing method.Experiments are per⁃formed on the 2D floorplan images.The results show the ef⁃fectiveness and applicability of the proposed vectorization method.