OPTIMIZATION OF 3D MODELS FOR USE IN AR ENVIRONMENTS

Abstract

The article presents a comprehensive theoretical analysis of modern approaches to optimizing three-dimensional models for use in Augmented Reality (AR) environments. The relevance of the study is determined by the rapid development of AR technologies in education, industry, medicine, e-commerce, and digital marketing, which is accompanied by increasing performance requirements for mobile and web-based platforms. One of the key challenges in implementing AR solutions is ensuring stable frame rates and fast loading of 3D content under limited computational resources of mobile devices. The purpose of this research is to systematize and generalize existing methods of 3D model optimization aimed at reducing polygon count, minimizing texture data volume, improving rendering efficiency, and ensuring compatibility with modern AR platforms. The paper analyzes mesh decimation algorithms, retopology techniques, Level of Detail (LOD) approaches, texture and normal map baking methods, texture atlases, graphical data compression, and visibility culling technologies. Special attention is given to optimization tools and software environments such as Blender, MeshLab, Simplygon, as well as AR engines including Unity and WebAR frameworks. The study demonstrates that effective optimization of 3D models requires a comprehensive combination of geometric, textural, and algorithmic techniques that significantly reduce model size without substantial loss of visual quality. It is proven that the use of multi-level object representation and adaptive rendering mechanisms enhances AR application performance, reduces loading time, and stabilizes frame rates. The obtained results can be applied in the development of high-performance AR systems within the framework of modern information technologies.