The focus on interior architecture within the real estate sector has intensified, with floor plans serving as the primary method for illustrating spatial layouts, encompassing rooms, doors, and windows. With advancements in image processing, computer vision, and graphics, it is now feasible to reconstruct and visualize building interiors in 3D. This paper concentrates on the approximate reconstruction of a suburban kindergarten in Sofia, Bulgaria, in 3D for future visualization of indoor pollution levels on a basis of data produced by sensors installed on the premises. We use well-established image processing algorithms for extracting features from a 2D floor plan, classify them statistically, and convert them into GeoJSON format for subsequent 3D reconstruction in Unity game engine. By successfully approximating the internal structure of the kindergarten building, our approach demonstrates significant potential. The results suggest avenues for refinement and application in comparable settings, offering advancements in building analysis and environmental monitoring alike.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

3D Reconstruction of Building Models Based on 2D Floor Plans Using Statistical Analysis

  • Pawel Gasiorowski,
  • Vassil Vassilev

摘要

The focus on interior architecture within the real estate sector has intensified, with floor plans serving as the primary method for illustrating spatial layouts, encompassing rooms, doors, and windows. With advancements in image processing, computer vision, and graphics, it is now feasible to reconstruct and visualize building interiors in 3D. This paper concentrates on the approximate reconstruction of a suburban kindergarten in Sofia, Bulgaria, in 3D for future visualization of indoor pollution levels on a basis of data produced by sensors installed on the premises. We use well-established image processing algorithms for extracting features from a 2D floor plan, classify them statistically, and convert them into GeoJSON format for subsequent 3D reconstruction in Unity game engine. By successfully approximating the internal structure of the kindergarten building, our approach demonstrates significant potential. The results suggest avenues for refinement and application in comparable settings, offering advancements in building analysis and environmental monitoring alike.