基于聚类算法和色彩网络的蝴蝶色彩分析及应用

蝴蝶蕴含着独特的自然美,是设计师重要的设计灵感来源。为厘清蝴蝶色彩及其搭配规律,更好地进行蝴蝶色彩重用,基于聚类算法和色彩网络对蝴蝶色彩进行分析和应用研究。首先,利用K-means聚类算法对单只蝴蝶8种不同角度的图像进行2次色彩聚类和主色提取,构建110只蝴蝶的6色标准色卡。其次,对蝴蝶标准色卡再次聚类,结合蝴蝶色彩实际搭配关系,构建了包含20个网络节点的蝴蝶色彩网络模型,并通过4个蝴蝶样本对色彩网络模型的可靠性进行验证。最后,以丝巾纹样配色设计作为应用案例,验证标准色卡和色彩网络模型作为配色辅助工具的可操作性和应用价值。案例表明,蝴蝶标准色卡和色彩网络模型可以为设计师在色彩设计过程中提供有效的辅助和支持。

Butterfly color analysis and application based on clustering algorithm and color network

The colorful butterfly with unique natural beauty is not only a common design theme, but also an important source of design inspiration. In order to clarify the color of butterfly as well as its matching rules and help designers better reuse the color of butterfly, the color of butterfly was studied and analyzed by clustering algorithm and color network. The color images of eight different angles of a single butterfly were firstly clustered twice using the K-means clustering algorithm, the main colors of butterfly were extracted, and the standard color cards of 110 butterflies were constructed. Then, the butterfly standard color cards were clustered for the third time, and the butterfly color network model including 20 network nodes was established, combined with the actual matching relationship of butterfly color. The reliability of the color network model was verified by six butterfly samples. Finally, the feasibility and application value of standard color card and color network model as color matching assistant tools were verified by the example of silk scarf pattern color matching design. Research and application cases show that butterfly standard color card and color network model can provide effective assistance and support for designers in the process of color design.