Image-based BIM texture generation method

Abstract

The invention discloses an image-based BIM texture generation method, and the method comprises the steps: carrying out the fusion of an obtained indoor image and an obtained BIM image, and obtaining a first image; processing the first image to obtain candidate textures of each semantic meaning; inputting the data of the model texture library as a data set into a texture generation model for training; and inputting the candidate texture of each semantic into a texture generation model to generate a plurality of candidate textures, evaluating the plurality of candidate textures, obtaining the candidate texture with the highest evaluation, and repairing the candidate texture to obtain the final BIM texture. According to the method, the BIM data is analyzed, the BIM is split into elements such as the ground, the wall, the ceiling and the table, and then the elements correspond to image recognition, so that correlation fusion of the BIM and image information is realized, textures are generated from pictures, the BIM information is fed back, and the method has the advantages of being independent of a texture library, high in efficiency and the like. The method has the advantages that dynamic updating can be achieved, and the corresponding relation between BIM construction and the monitoring image can be obtained without depending on training semantics.

Description

technical field [0001] The invention belongs to the field of BIM texture generation, in particular to an image-based BIM texture generation method. Background technique [0002] At present, digital twins have become a research hotspot and frontier in academia and industry. BIM, as a high-precision digital model, is an important basis for digital twins. However, BIM only has structural and semantic information, and textures are often in CAD format, only a single color , such a digital model that is not realistic enough cannot meet the needs of digital twins. Surveillance video is the most easily obtained and the largest source of image information. It records real texture information and fuses it with BIM information to extract real texture information and give it to the corresponding components in BIM. This will make BIM more real and also Can lay the groundwork for a true digital twin. [0003] The main purpose of texture reasoning integrating BIM and computer vision is t...

Images