EngEnglish language
PERM TECH SCIENTISTS TRAINED ARTIFICIAL INTELLIGENCE TO RECOGNIZE THE IMAGE OF THE CITY

PERM TECH SCIENTISTS TRAINED ARTIFICIAL INTELLIGENCE TO RECOGNIZE THE IMAGE OF THE CITY

23 July 2021 127

Artificial intelligence and machine learning technologies are being actively implemented today in architecture and urban planning. They help to optimize planning solutions, engineering systems, as well as provide safe and comfortable living conditions for people. But there are no analogues of intelligent decision support systems for the color design of facades in Russia and in the world so far.

— We trained the system to use digital images of buildings. Intelligent analysis helped to find out what colors are typical for certain cities, and it can find out which settlements are similar to each other. The development allows to determine which city it belongs to using photograph of a residential complex only, - explains Olga Bolotova, a graduate of the Perm Tech Civil Engineering Faculty.

The objects of the study were apartment buildings in the Russian largest cities with a population of over one million. The intelligent solution will become a recommendation system for architects and specialists in related fields. According to the developers, in the creative process of searching for the color design of a building, a person cannot be replaced by it, but the system can become a digital assistant.

The researchers analyzed 15 images of residential complexes under construction in Perm that were not used during the training process. The system determined that all the houses are new buildings, and the complexes are similar to Perm. The development identified 13 buildings as Perm, and two residential complexes, according to the results of the analysis, turned out to be similar to other cities: Yekaterinburg and Omsk. For one of them, the developer identified the most relevant Perm colors and proposed options for changing the facades.

—The software solution can be used by developers to audit conceptual design. The system will help to give recommendations on the building's design so that they meet the goals set and can be fit into the urban space, - explains Irina Alekseeva, the project manager, associate professor of the Construction Engineering and Materials Science Department of the Perm Tech, Ph.D. in Economics.

To create the development, in particular, for the spectral analysis of images and identification of the information models parameters, the researchers used domestic software.

Olga Bolotova's intellectual system has already been highly appreciated by the  Urban Planning and Architecture Department of the Perm Administration. The developers are planning to create similar solutions for buildings of general education and preschool educational institutions, sports and recreation centers, swimming pools and other public buildings, as well as to study color patterns in an urban environment to check the uniformity of the design of buildings on the same territory. Researchers are engaged in these areas on the recommendation of the chief architect of the city, Dmitry Lapshin.