High-Fidelity Dynamic Texture Rendering Method with Engineering Feature Awareness

In digital twins, engineering visualization, and virtual reality applications, the quality of texture rendering not only determines visual realism but also directly affects the expression and analysis of engineering information. Details such as nameplates, welds, cracks, and wear in complex scenes—such as railways, architecture, and machinery—serve as critical references for engineering decision-making and maintenance. However, traditional rendering methods often struggle to simultaneously ensure both high-fidelity texture details and real-time system performance in large-scale scenarios. To address this challenge, this paper proposes a high-fidelity dynamic texture rendering framework that incorporates engineering feature awareness. Centered on engineering feature perception, the method integrates multi-resolution geometry and texture hierarchy management, feature-sensitive LOD strategies, frustum culling with GPU parallel optimization, as well as deferred rendering and controlled ray tracing. The framework aims to maintain the recognizability of engineering features while achieving real-time performance. Furthermore, this paper systematically analyzes the theoretical characteristics, experimental design considerations, and application prospects of the proposed method, thereby providing a technical roadmap and research reference for future implementation and validation.