基于高分辨率卫星影像数据高原输电通道巡视技术的研究

近年来，随着西电东输工程的持续推进，位于高原区域的输电线网规模持续扩大，但因受高原气候条件影响，高原区域内的输电通道巡视难度逐步增大。为提升高原输电通道的巡视精度、效率，保障高原输电通道的安全，本文以worldview-3高分卫星影像数据为数据源，采用多尺度分割技术与面向对象分类算法，研究分析高原输电通道关键地物信息的提取方法，研究表明：（1）面向对象算法适用于高分影像数据的信息提取，其地物信息提取精度较高，能够满足输电通道的巡视工作，这表明基于高分卫星影像技术的输电通道巡视可行；（2）基于多尺度分割技术耦合面向对象分类算法提取的大棚信息精度可达95%，可替代人工解译工作，而道路信息精度可相对较低，但仍可替代大部分人工解译工作；（3）与传统野外核查、人工目视解译相比，基于多尺度分割技术耦合面向对象分类算法具有边界精度高、效率高等优势。

Research of high altitude transmission channel inspect technology based on high resolution satellite image data

In order to improve the inspection accuracy and efficiency of the plateau power transmission channel and ensure the safety of the plateau power transmission channel, this paper uses worldview3 highresolution satellite image data as the data source, adopts multiscale segmentation technology and objectoriented classification algorithm to study and analyze the extraction method of the key ground feature information of the plateau power transmission channel. The research shows that:(1)the objectoriented algorithm is suitable for the information of highresolution image data.The accuracy of ground feature information extraction is high, which can meet the inspection work of transmission channel and show that the inspection work of transmission channel based on highresolution satellite image technology is feasible.(2)The accuracy of canopy information extraction based on multiscale segmentation technology coupled with objectoriented classification algorithm can reach 95%, which can replace the manual interpretation work, while the accuracy of road information can be relatively low, but it can still replace most of the work.(3)Compared with the traditional field inspection and artificial visual interpretation, the multiscale segmentation based objectoriented classification algorithm has the advantages of high boundary accuracy and high efficiency.