To address issues such as unreasonable band selection and feature redundancy in waste textile detection, this study proposes a multi-attribute group decision-making framework based on cubic q-rung orthopair fuzzy sets for optimizing infrared bands. Firstly, a novel cubic q-rung orthopair fuzzy power Maclaurin symmetric mean (Cq-ROFPMSM) operator is developed by integrating the PA and MSM operators to capture correlations among spectral attributes and mitigate the influence of outliers, thereby enabling more effective fusion of uncertain spectral information. To objectively determine attribute weights, a novel hybrid weighting MEREC-CRITIC method was proposed, which can reflect both the mutual influence among attributes and their overall importance. With five typical waste textiles as decision objects, the information content, separability, and correlation of the bands are selected as evaluation attributes. Finally, the CODAS method is employed to rank five band division schemes, and the results indicate that the proposed model identifies an optimal band range of 1800 ~ 2600 nm. Compared with the baseline methods such as KNN, DT and CNN, the proposed method improves image quality by 35% and classification accuracy by 25% in waste textile detection. This study contributes a theoretically robust and practically applicable decision framework for spectral band optimization in waste textile recycling.
Zhao et al. (Tue,) studied this question.