基于多源数据与K-means聚类的地域特征关联纸质包装色彩应用研究

Application Research of Regional Feature Correlation Packaging Color Based on Multi-source Data and K-means Clustering

关键词:  中央大街  地域特征  历史街区  纸质包装

Key Words:Zhongyang Street  regional characteristic  historic district  paper packaging

摘要:以哈尔滨中央大街景区为研究对象，采用OSM开源地图平台、百度API平台配合ArcGIS10.6平台、Python平台获取数据，通过Matlab平台下DeepLabV3+模型对获得的街道图像进行语义分割，通过K-means聚类算法以HSB色彩空间为模型基准，绘制色彩散点聚类模型、色彩连续性模拟模型与片段色彩聚类模型对中央大街景区进行可视化分析，并提出地域特征关联纸质包装设计的色彩应用建议。多源数据结合K-means聚类法可以在短时间内相对快速、精准、科学完成纸质包装色彩提取与应用策略制订，其结果能够体现地域文脉特色，提升地域特色包装辨识度，有助于实现地域特征关联纸质包装色彩方案优选，为纸质包装色彩应用提供参考。

Abstract:Taking the scenic spot of Harbin Zhongyang Street as the research object， OSM open source map platform， Baidu API platform， ArcGIS10.6 platform and Python platform were used to obtain data. DeepLabV3+ model under Matlab platform was used to perform semantic segmentation on the obtained street images. Using the HSB color space as the model benchmark， the K-means clustering algorithm drew the color scatter clustering model， color continuity simulation model and fragment color clustering model to visualize and analyze the Zhongyang Street scenic spot， and put forward the color application suggestions of regional characteristics associated paper packaging design. The multi-source data combined with K-means clustering method could complete the paper packaging color extraction and application of strategy formulation relatively fast， accurately and scientifically in a short period of time. The results reflected the regional cultural packaging local characteristics， enhanced the recognition of regional characteristics packaging which helped to achieve regional characteristic associated paper packaging color scheme optimization， and provided reference for paper packaging color application.