Method for Constructing a Facade Dataset through Deep Learning-Based Automatic Image Labeling

The construction industry has made great strides in recent decades by utilizing computer programs, including computer aided design programs. However, compared to the manufacturing sector, labor productivity is low because of the high proportion of knowledge-based tasks and simple repetitive tasks. Therefore, knowledge-based task efficiency should be improved through the visual recognition of information by computers. A computer requires a large amount of training data, such as the ImageNet project, to recognize visual information. This paper proposes facade datasets that are efficiently constructed by quickly collecting facade data through road-view images generated from web portals and automatically labeled using deep learning as part of the construction of image datasets for visual recognition construction by a computer. Therefore, we attempted to automatically label facade images to quickly generate large-scale facade datasets with much less effort than the existing research methods. Simultaneously, we constructed datasets for a part of Dongseong-ro, Daegu Metropolitan City, and analyzed their utility and reliability. It was confirmed that the computer could extract significant facade information from the road-view images by recognizing the visual information of the facade image. In addition, we verified the characteristics of the building construction image datasets. This study suggests the possibility of securing quantitative and qualitative facade design knowledge by extracting facade design information from facades anywhere in the world. Previous studies mainly collected facade images through camera photography to construct databases, but in this study, a significant part of the database construction process was shortened through automation. In the case of facade automatic image labeling studies, it is the facade-based automatic 3D modeling which has been primarily studied, but it is difficult to find a study to extract data for facade design research.

Automated summary

The construction industry has seen great progress in recent decades due to the use of computer programs, but labor productivity remains low compared to the manufacturing sector. To improve the efficiency of knowledge-based tasks, this study proposes a method to construct efficient facade datasets by collecting road-view images and automatically labeling them using deep learning. The study confirms that computers can extract significant facade information from images and verifies the characteristics of the building construction image datasets. By extracting facade design information from facades worldwide, this study suggests the possibility of securing both quantitative and qualitative facade design knowledge. The automation in the database construction process has shortened the overall time required.