Dynamic drape indicator extraction for fabrics based on Fourier-RF projection-contour decomposition

To stably characterize projected-area fluctuation and nonaxisymmetric lobelike contours in dynamic drape testing of circular fabrics, this study proposes a Fourier-RF projection-contour decomposition method. The method establishes a physically interpretable dynamic drape descriptor system that separates area-scale response, low-order asymmetric disturbance, primary lobe modes, and high-order residual irregularity from projection videos. After removing the acceleration stage, projection masks are obtained by background separation, connected-component filtering, hole filling, and morphological closing, and frame-wise outer-contour radius functions are extracted in polar coordinates. Fourier-RF layered representation is compared with equivalent circle, Chebyshev basis, principal component analysis (PCA)/proper orthogonal decomposition (POD), and Fourier modal forms, and a dynamic drape indicator set is constructed from Fourier-RF energy decomposition, including area contraction, area fluctuation, dominant-frequency response, asymmetric disturbance, lobe-mode intensity, modal order, modal entropy, and random-field residual. For 80 fabric-speed cases, deterministic Fourier 1–8 representation achieves a mean relative root-mean-square error (RMSE) of 0.015, lower than the Chebyshev baseline (0.044) and equivalent circle (0.069); adding a high-order random-field residual further reduces it to 0.006. The indicators distinguish dynamic drape states with similar area contraction but different lobe-mode spectra. Forward association shows modest correlations of SDDC and DAFI with material parameters and speed (0.210 and 0.253). In finite-element method (FEM) forward simulation, compared with the Chebyshev angular field, the Fourier-RF angular field reduces SDDC mean absolute error (MAE) from 0.426 to 0.231 and DAFI MAE from 1.107 to 0.491. These results indicate that Fourier-RF projection-contour indicators capture lobelike morphological complexity beyond traditional area indicators and provide a stable, physically interpretable system for dynamic drape evaluation.