基于可见光光谱的装饰纸表面视觉特征的模式识别研究

发布时间：2018-04-20 19:32

本文选题：装饰纸 + 可见光光谱　；参考：《光谱学与光谱分析》2015年02期

【摘要】：装饰纸作为现代家居产品中必不可少的重要饰面材料之一,不仅具有良好的装饰效果,还可以大大地改善材料的表面性能。然而,装饰纸外观质量的控制(如色差的评价和控制),仍然是制约装饰纸产业发展的重要因素之一。利用人眼来判别和控制往往因人而异,偏主观而非客观。因此,寻找一种客观、有效的方法来代替人的视觉感官来快速判别具有重要意义。本研究利用可见光光谱结合主成分分析法对不同类别装饰纸进行了模式识别分类研究,探讨了该技术的可行性。结果表明:(1)可见光光谱与表征装饰纸表面视觉特性的参数之间呈现出显著的相关性,相关系数都达到了0.85以上,多数均达到0.99,说明可见光光谱中包含有描述装饰纸表面视觉特征的参数;(2)利用装饰纸表面的可见光光谱结合主成分分析方法所建立的模式识别模型,对不同装饰纸表面视觉特征类型进行模式识别时,正确率达到94%~100%,说明可见光光谱分析技术有潜力成为装饰纸表面视觉特征快速、客观、有效识别和分类的新技术。
[Abstract]:As one of the most important decorative materials in modern household products, decorative paper not only has good decorative effect, but also greatly improves the surface properties of materials. However, the control of decorative paper appearance quality, such as color difference evaluation and control, is still one of the important factors restricting the development of decorative paper industry. The use of human eyes to distinguish and control often vary from person to person, subjective rather than objective. Therefore, it is of great significance to find an objective and effective method to replace human visual senses. In this study, the pattern recognition classification of different decorative paper was studied by means of visible light spectrum and principal component analysis, and the feasibility of this technique was discussed. The results show that there is a significant correlation between the visible light spectrum and the parameters that characterize the visual properties of decorative paper surface, and the correlation coefficient is above 0.85. Most of them are 0.99, indicating that the visible light spectrum contains the parameter of describing the visual characteristics of the decorative paper surface, and the pattern recognition model is established by using the visible light spectrum of the decorative paper surface and the principal component analysis (PCA) method. The correct rate of pattern recognition for different types of visual features of decorative paper surfaces is 94 / 100, which indicates that visible light spectrum analysis has the potential to become a new technique for recognizing and classifying the visual features of decorative paper surfaces quickly, objectively and effectively.
【作者单位】：中国林业科学研究院木材工业研究所;
【基金】：国家自然科学基金项目(30800889,31370711) 国家人力资源和社会保障部留学回国人员科技择优优秀项目资助
【分类号】：O657.3;TS77

【参考文献】