This study takes digital modeling technology as the core and explores its application in the innovative design of traditional ethnic clothing. By integrating particle dynamics model, triangular mesh delineation algorithm and dynamic feature point co-tracking technology, three major modules of fiber fabric simulation, garment-human body motion co-tracking and immersive virtual interaction system are constructed. Experimental validation shows that the fabric simulation method based on 3D mesh convolution outperforms the traditional physical simulation and deep learning model in terms of Hausdorff distance error (2.381 mm for flag, 5.318 mm for teapot, 5.343 mm for hanging cloth) and time consumption (3.62 s for flag, 2.84 s for teapot, 3.31 s for hanging cloth), and the efficiency of vertex extraction is improved by 40.5%, which significantly reduces the computational cost. Factor analysis further extracted the three core evaluation factors of “fabric characteristics”, “sample structure” and “dynamic effect” (cumulative variance contribution rate of 90.17%). The results of subjective evaluation by users show that the page layout (4.35~4.88 ratings by ordinary users and 4.78~4.94 ratings by administrators) and the comprehensiveness of functions (4.27~4.84 ratings by ordinary users and 4.57~4.88 ratings by administrators) of the virtual interactive display system are highly recognized. The research results provide highly efficient technical support for the digital heritage and innovative design of ethnic costumes, with both theoretical and applied values.
