基于熵权TOPSIS和图像处理的保暖纤维保暖能力的评估

Evaluation of Thermal Insulation Capability of Thermal Fiber Based on Entropy Weight TOPSIS and Image Processing

摘要: 冬季服装保暖依赖填充材料，故对保暖纤维的保暖能力定量刻画十分重要。本文应用t-SNE算法和熵权TOPSIS模型构建保暖纤维指标体系，对保暖能力指标降维、分配熵权权重并排名，结果表明羊绒的保暖性能最优。然后，采用岭回归模型分析保暖性能与纤维平均长度和直径的关系，得出增加长度与表面积密度、减小直径可提高保暖性能。进一步，选取纤维直径估测部分纤维保暖能力，用Python处理部分纤维图像。最后，从宏观与微观层面评估保暖纤维，为其性能评估及新材料开发提供参考。

Abstract: Winter clothing insulation relies on filling materials, so it is important to quantitatively characterize the insulation ability of thermal fibers. This article applies t-SNE algorithms and entropy weight TOPSIS models to construct a thermal fiber index system, reduces the dimensionality of thermal insulation capacity indicators, assigns entropy weight and ranks them, and the results show that cashmere has the best thermal insulation performance. Then, the ridge regression model was used to analyze the relationship between thermal insulation performance and fiber average length and diameter, and it was found that increasing length and surface area density, and reducing diameter can improve thermal insulation performance. Furthermore, the fiber diameter is selected to estimate the thermal insulation capacity of some fibers, and the part of fiber images are processed by Python. Finally, evaluate the insulation fibers from both macro and micro perspectives to provide reference for their performance evaluation and new material development.