Evaluating the ability of deep learning to predict urban growth and expansion (limited case study: Isfahan city and suburbs)

taherebakhtiari¹ mahdiyehmorsali²

1) M.Sc. in Remote Sensing and GIS, Kharazmi University, Tehran, Iran.
2) Associate Degree in Urban Architecture, Hooshmand Sazeh University of Applied Science and Technology, Tehran, Iran.

IOI Code : ACCD-ACAAKA
Publication : 10th International Conference on Applied Researches in Science & Engineering (carseconf.com)
Abstract :
    Urban expansion is one of the fundamental challenges in urban management and planning and achieving sustainable development. Identifying and analyzing the factors affecting this process plays an important role in understanding urban growth patterns and supporting long-term decision-making. In this regard, deep neural networks have a significant capacity to analyze and predict urban land use changes due to their ability to model complex and nonlinear relationships between variables. The present study aimed to analyze the expansion trend of urban areas in Isfahan and its surrounding areas during the years 1990 to 2020 and evaluate the performance of deep neural networks in modeling these changes. For this purpose, land use data in the study period and spatial and environmental variables including distance from the road, slope, elevation, geology, and soil were used as model inputs. The collected data were used to train and evaluate the model to examine the network s ability to identify urban expansion patterns based on the learned relationships between input variables. The evaluation results showed that the developed model achieved a value of 0.95 for the accuracy criterion and 0.2 for the error criterion. These results indicate the model s ability to represent urban expansion patterns in the study area and can provide a basis for analyzing spatial changes and supporting urban development planning and management in Isfahan and its surrounding areas.

Keywords : ability assessment deep learning urban growth prediction isfahan city and suburbs