APPLICATION OF GIS AND ARTIFICIAL INTELLIGENCE TECHNOLOGIES FOR VEGETATION IDENTIFICATION

Abstract

The integration of artificial intelligence (AI) methods - machine and deep learning - into GIS has created the prerequisites for a new field of GeoAI, the goal of which is to accelerate spatial decision - making, reduce the need for fieldwork, increase the accuracy of data processing, and reduce costs. Therefore, research focused on the application of advanced AI models in the GIS environment and their practical applicability for creating automated digital surfaces is extremely relevant both in the context of science and practical implementation.

In urbanized areas, the terrain and structure of objects are complex and very dense. This poses challenges in accurately and automatically identifying buildings, vegetation, and determining their real heights. Identification of urban greenery (trees, shrubs) is a very important component of sustainable urban planning, which is associated with microclimate, air quality, stormwater regulation, and ecosystem assessment. Traditional greenery inventories are expensive and rarely updated, therefore remote sensing data and AI analysis allow to significantly reduce the costs of measurements and increase the periodicity of updates.

The article examines the synergy of applying geographic information systems (GIS) and artificial intelligence (AI) for vegetation identification in an urbanized area using digital spatial laser scanning point data and orthophotographic images of county centers of the Republic of Lithuania. During the study, it was determined that when applying GIS and AI technologies for vegetation identification, the difference in the area occupied by tree foliage amounted to 0.0249 ha 194 trees were identified, the area occupied by foliage amounted to 0.486 ha. The identified vegetation constituted 17.6 percent. of the entire territory under study.