一种影像特征挖掘的无监督道路快速提取方法OA
An unsupervised fast road extraction method based on image feature mining
作为基础地理信息数据的重要组成部分,道路的自动化提取一直是遥感监测领域的关键课题.传统人工识别效率低且耗时长,本文基于聚类分析技术,开展了遥感影像道路信息检测的探索.先对遥感图像进行分割处理,再实施聚类分析,结合边缘检测算法和道路搜索判定算法来快速提取道路信息.研究结果表明:本文道路提取方法综合计算值F1高于80%,单幅图像的处理时间为2~6 s,能够高效提取道路边缘,可在测绘生产中进行辅助作业,大幅提高工作效率.
Roads constitute a crucial component of fundamental geographic information data,and their automated extraction has long been a key research topic in the field of remote-sensing monitoring.Traditional manual recognition is inefficient and time-consuming.Therefore,this study explored road information detection in remote-sensing images based on cluster analy-sis technology.First,the image was segmented;then clustering was performed,combined with an edge detection algorithm and a road-search decision rule to rapidly extract road information.The results demonstrate that the proposed method achieves a comprehensive F1-score above 80%,and the processing time for a single image ranges from 2 to 6 s.It effectively and rapidly extracts road edges,making it feasible as an auxiliary tool in surveying and mapping production and capable of substantially improving work efficiency.
马红;陈浩
广东省国土资源测绘院,广东 广州 510700||自然资源部华南热带亚热带自然资源监测重点实验室,广东 广州 510700广东省国土资源测绘院,广东 广州 510700||自然资源部华南热带亚热带自然资源监测重点实验室,广东 广州 510700
天文与地球科学
聚类分析道路提取智能识别边缘检测测绘生产
cluster analysisroad extractionintelligent identificationedge detectionsurveying and mapping production
《北京测绘》 2026 (9)
1225-1231,7
广东省省级科技计划(2021B1212100003)广东省自然资源厅科技项目(GDZRZYKJ2023001).
评论