测绘通报 ›› 2021, Vol. 0 ›› Issue (4): 105-110,125.doi: 10.13474/j.cnki.11-2246.2021.0119

• 技术交流 • 上一篇    下一篇

多山地区输电通道内道路提取方法

彭赤1, 毛先胤2, 杨柳青2, 马晓红2, 吕乾勇2, 杜昊1, 王迪3, 余容3   

  1. 1. 贵州电网有限责任公司生产技术部, 贵州 贵阳 550000;
    2. 贵州电网有限责任公司电力科学研究院, 贵州 贵阳 550000;
    3. 中国电建集团贵州电力设计研究院有限公司, 贵州 贵阳 550000
  • 收稿日期:2020-06-03 出版日期:2021-04-25 发布日期:2021-04-30
  • 作者简介:彭赤(1976-),男,高级工程师,研究方向为输电线路建设及运行检修。E-mail:2082770019@qq.com
  • 基金资助:
    贵州电网有限责任公司科技项目(GZKJXM20181775)

Road extraction method in transmission channel in mountainous area

PENG Chi1, MAO Xianyin2, YANG Liuqing2, MA Xiaohong2, Lü Qianyong2, DU Hao1, WANG Di3, YU Rong3   

  1. 1. Production Technology Department, Guizhou Power Grid Co., Ltd., Guiyang 550000, China;
    2. Electric Power Research Institute of Guizhou Power Grid Co., Ltd., Guiyang 550000, China;
    3. Power China Guizhou Electric Power Design & Research Institute Co., Ltd., Guiyang 550000, China
  • Received:2020-06-03 Online:2021-04-25 Published:2021-04-30

摘要: 准确提取输电通道内各地物类别是进行输电通道三维建模及交跨分析的基础。为解决目前贵州等多山地区输电通道内道路提取方法稳健性低、效率低的问题,本文通过结合道路特征显著性检测与形态约束的方法对影像中多山地区输电通道内道路的快速准确提取进行了研究。首先,对卫星影像中山区道路特征进行分析,对每个像素分别计算基于道路颜色一致性和结构一致性的像素级显著值;然后,结合显著目标空间先验知识融合显著性检测结果,形成最终道路显著图,初步提取影像中道路目标;最后,分析道路与建筑物等的差异性,基于道路形态一致性定义道路形态约束条件,通过制定约束规则最终实现道路的准确提取。结果表明,该方法对于不同弯曲程度、粘连程度及影像对比度的道路都能实现快速准确提取,提取完整度、正确率、质量及耗时平均分别为97.5%、97.0%、95.6%、0.515 s。该方法稳健性高,可以快速、准确提取各种情况下道路,在输电线路工程实际中有很好的应用前景。

关键词: 多山地区输电通道, 道路提取, 结构一致性, 颜色一致性, 形态一致性, 像素级显著性, 形态约束

Abstract: Accurate extraction of objects in transmission channels is the basis of three-dimensional modeling and intersection analysis of transmission channels. In order to solve the problems of low robustness and low efficiency of road extraction methods in transmission channels in mountainous areas such as Guizhou, in this paper, the fast and accurate extraction of roads in transmission channels in mountainous areas in images is studied by combining road feature significance detection and morphological constraints. First of all, the road features in the mountain area in the satellite image are analyzed, and the pixel-level significant values based on road color consistency and road structure consistency are calculated for each pixel. And then, combined with the a priori knowledge of significant target space, the saliency detection results are fused to form the final road saliency map, and the road targets in the image are extracted preliminarily. Finally, the differences between roads and buildings are analyzed, the road shape constraints are defined based on the consistency of road shape, and the accurate extraction of roads is realized through the formulation of constraint rules. The results show that the roads with different bending degree, different adhesion degree and different image contrast can be extracted quickly and accurately, and the average extraction integrity, accuracy, quality and time-consuming are 97.5%, 97.0%, 95.6% and 0.515 s, respectively. It can be seen that this method has high robustness, can extract roads in various cases quickly and accurately, and has a good application prospect in transmission line engineering.

Key words: transmission channel in mountainous area, road extraction, structural consistency, color consistency, morphological consistency, pixel significance, morphological constraint

中图分类号: