Building facade texture automatic generation method and system

Abstract

The invention relates to the technical field of computer vision, in particular to a building facade texture automatic generation method and system. According to the method, a deep learning network capable of carrying out object identification is adopted to carry out window extraction, a building facade texture is synthesized by using a Pix2PixHD network for synthesizing a new building facade wall texture, and finally, a window is pasted on a synthesized building facade texture image according to obtained window information and a real layout structure of the window information. According to the method, the real building facade texture is generated by adopting image generation and window extraction methods, and the real layout information and the actual size of the window are kept according to the extracted mask file of the window, so that the time for manually obtaining the position and the size of the window is saved; the generated building facade keeps the real position and structural layout of the window on the original facade, ensures the authenticity of the layout of the object on the facade, and generates a building facade image which is more in line with the facade size in an actual model.

Description

technical field [0001] The invention relates to the technical field of computer vision, to computer vision image generation, three-dimensional city modeling, and digital twin technology, and in particular to a method and system for automatically generating building facade textures based on multi-view image modeling data. Background technique [0002] The analysis and three-dimensional modeling of real urban scenes are the basic problems of computer vision and computer graphics. With the wide use of three-dimensional urban architectural models in the fields of house survey, stereo measurement and urban planning, the construction of three-dimensional models of buildings received great attention. Due to the advantages of low modeling cost, rich model texture details, short modeling time and fast speed, oblique photographic modeling has now developed into a mainstream method for large-scale reconstruction. [0003] At present, most oblique photographic modeling adopts the multi...

AI Technical Summary

Problems solved by technology

[0005] In order to solve the problem that some building facades in urban 3D modeling are prone to distortion and poor texture quality, and it is difficult to meet the requirements of 3D buildings to express real scenes, the purpose of the embodiments of the present invention is to provide a building facade texture automatic The generation method and system focus on the problem of automatic replacement of building facade textures. By replacing distorted and hollow building facades, the purpose of complete building facades is achieved. The buildings after facade replacement can make three-dimensional The buildings in the model look more regular and unified, and the overall model looks neater

Images