Objective To address the issues of cultural distortion and insufficient controllability in the intelligent generation of traditional patterns, this study proposes a parameter-constraint-guided intelligent design method using Pan Yao patterns as a case study. The approach aims to achieve precise preservation of cultural elements and controllable innovation in traditional patterns through digital constraints during the intelligent design process.

Methods A four-stage methodology—"cultural analysis-parameter constraints-intelligent generation-practical validation"—was employed. During pattern analysis, field research and image analysis are used to extract semantic, structural, and chromatic genes that represent ethnic identity markers from Pan Yao patterns, thus establishing a tripartite constraint system integrating "semantics-structure-color". During pattern processing, constrained by semantic logic, three sets of pattern primitives were selected. By integrating Pan Yao's compositional principles with mathematical algorithms and using tessellation geometry as the foundation, the pattern primitives were configured and arranged in Grasshopper, which achieved rational derivation of the patterns while preserving their cultural characteristics. During pattern generation, to validate the feasibility of intelligent parameter-constrained generation, a non-parameter-constrained control group was established. The ControlNet module of the Stable Diffusion platform was employed to achieve color filling and material simulation of the patterns. For color control, the K-Means clustering algorithm was employed to extract patterns' hues, while text prompts guided the color filling process. Material simulation utilized the Flux-dev model as the core architecture, achieving textured reproduction through systematic parameter configuration and multi-module coordination. During the pattern application and evaluation stage, a women's leather handbag served as the carrier. NURBS technology that was integrated with Grasshopper enabled surface mapping of the patterns onto the handbag. A multi-stakeholder evaluation panel comprising 20 members, including relevant scholars and designers, along with 15 female consumers, conducted a comprehensive assessment of the design proposals using the Fuzzy Comprehensive Evaluation (FCE) method.

Results A comparative experiment without parametric constraints revealed that intelligently generated patterns under parametric constraints demonstrated significant advantages in hierarchical richness, cultural alignment, and visual impact. This approach effectively mitigates critical issues in traditional intelligent pattern generation, such as cultural distortion and lack of controllability. Simultaneously, the parametric modeling approach utilizing NURBS technology, combined with the Grasshopper visual programming environment, enables precise pattern mapping onto complex surfaces. Evaluation by a multi-stakeholder panel comprising Yao ethnic costume scholars, designers, and consumers indicated that Product 1 from Scheme 1 and Product 6 from Scheme 2 achieved the highest scores across comprehensive metrics, including pattern morphological similarity, semantic heritage preservation, craftsmanship feasibility, and market acceptance. Their satisfaction rates reached 80.15% and 81.65%, respectively, validating the effectiveness of this design methodology.

Conclusion This study demonstrates that the three-stage design methodology—parametric constraints, intelligent generation, and feasibility validation—developed using Pan Yao patterns as a case study achieves a dynamic equilibrium of precision, controllability, and richness in intelligent pattern generation. This is accomplished through the synergistic application of explicit rule-based constraints and modular design strategies. This system effectively addresses critical issues in traditional intelligent pattern generation, such as cultural semantic decay, morphological distortion, and uncontrollable generation processes, thereby partially overcoming the limitations of experience-driven design models. It provides a replicable and scalable theoretical framework and practical pathway for empowering the innovative development of ethnic minority costume patterns through digital and intelligent technologies.