遥感信息2024,Vol.39Issue(3) :90-96.DOI:10.20091/j.cnki.1000-3177.2024.03.013

融合边缘优化SLIC和A*算法的影像镶嵌线提取

Image Mosaic Line Extraction by Fusion of Edge Optimization of SILC and A* Algorithm

常巧梅 张浩 张娟
遥感信息2024,Vol.39Issue(3) :90-96.DOI:10.20091/j.cnki.1000-3177.2024.03.013

融合边缘优化SLIC和A*算法的影像镶嵌线提取

Image Mosaic Line Extraction by Fusion of Edge Optimization of SILC and A* Algorithm

常巧梅 1张浩 2张娟3
扫码查看

作者信息

  • 1. 山西工程职业学院,太原 030032
  • 2. 中国科学遥感与数字地球研究所,北京 100094
  • 3. 山西亚太遥感新技术有限公司,太原 030032
  • 折叠

摘要

针对遥感影像自动提取镶嵌线穿过独立地物损害地物完整性的问题,提出融合边缘信息优化SLIC和A*算法的影像镶嵌线提取方法.首先,提取影像边缘信息,并将边缘强度因子与影像光谱信息融合;然后,使用优化的简单线性迭代聚类方法对重叠区域影像进行超像素分割,获取地物完整的超像素,并利用数学形态学算子去除孤立噪声与不完整边界;最后,采用曼哈顿距离启发函数提升A*算法的镶嵌线提取效率,获取最优的镶嵌线提取.分析结果表明,该方法可有效提取影像镶嵌线,避免镶嵌线通过独立建筑物,满足后期正射影像制作要求.

Abstract

In order to solve the problem of automatic extraction of remote sensing image mosaic lines which pass through independent features and damage the integrity of features,an image mosaic line extraction method combining edge information optimization SLIC and A* algorithm is proposed.First,the edge information is extracted and the edge intensity factor is fused with the spectral information.Then,an optimized simple linear iterative cluster is used to implement superpixel segmentation to obtain complete superpixels of ground objects,and mathematical morphology operators are used to remove isolated noises and incomplete boundaries.Finally,Manhattan distance heuristic function is used to improve the efficiency of mosaic line extraction of A* algorithm,which obtains the optimal mosaic line extraction.The experimental comparison and analysis of remote sensing images show that the proposed method can effectively extract mosaic lines,avoid mosaic lines passing through independent buildings,and meet the requirements of later orthophoto production.

关键词

边缘信息/简单线性迭代聚类/A*算法/镶嵌线提取

Key words

edge information/simple linear iterative clustering/A* algorithm/mosaic line extraction

引用本文复制引用

基金项目

国家自然科学基金(41771397)

出版年

2024
遥感信息
科学技术部国家遥感中心,中国测绘科学研究院

遥感信息

CSTPCDCSCD北大核心
影响因子:0.712
ISSN:1000-3177
参考文献量13
段落导航相关论文