摘要
根据无人机影像电力线的特征,利用LSD算法实现了电力线的自动提取。首先使用各向同性的Sobel算子对特征区域进行图像增强,再使用LSD算法进行线段提取,然后通过线段合并和线段拟合过滤线段集合,最后使用K均值算法进行聚类,筛选得到最终电力线结果。采用多种地物背景下的电力线影像进行实验,结果表明,该方法能提取完整的电力线,稳健性强,对电力巡线具有较大的实际意义。
In this thesis, according to the features of UAV image powerline, we proposed a powerline extraction method based on LSD algorithm. Firstly, we used horizontal Sobel operator to enhance the edge in the area, and then used LSD algorithm to extract line set. Finally, we used the line angle and line position to filter line set, and adapted K-means algorithm for clustering to determi ne the result. The experimental result shows the method can extract the powerline exactly under different background.
出处
《地理空间信息》
2019年第1期35-37,71,123,共5页
Geospatial Information
