众包图像数据采集与聚类分析方法探讨  

Data Acquisition and Clustering Analysis Method of Crowd Sourcing Images

作　　者：余晓敏[1] 陈尔刚 季鹏 郭涛[2] 秦昆[2] 

机构地区：[1]湖北省基础地理信息中心,湖北武汉430074 [2]武汉大学遥感信息工程学院,湖北武汉430079

出　　处：《地理空间信息》2017年第11期16-17,20,共3页Geospatial Information

基　　金：国家科技支撑计划课题资助项目(2015BAJ05B01)

摘　　要：众包图像是由大众经过一定方法获取后通过互联网向公众或相关机构提供的一种开放式图像数据。利用网络爬虫工具在互联网上爬取了一定数量的众包图像,并分别探讨了单张图像聚类方法和多张图像聚类方法,以期为众包技术如何服务于智慧小城镇规划管理提供技术参考。利用K-means聚类方法对单张众包图像进行聚类,并探讨了分别利用Python语言和Java语言编程实现图像聚类的方法 ;利用层次聚类方法对多张众包图像进行聚类。Crowd sourcing images are a kind of open sourcing image data, which are collected and provided to citizens or organizations through Internet. In this paper, the crawler technology was used to crawl crowd sourcing images from Internet, and the image clustering analysis methods were explored for single image and multiple images separately. The K-means clustering method was used to cluster single image, which was programmed by Python and Java language respectively. And the hierarchical clustering method was used to cluster multiple images. This study can provide some supports for the planning and management of smart small towns.

关 键 词：众包图像 智慧小城镇 数据采集 图像聚类 K-MEANS聚类 层次聚类 

分 类 号：P237[天文地球—摄影测量与遥感]