基于特征检测的航拍图像电力线提取方法  

Power line extraction from aerial images based on feature detection algorithm

作　　者：张少平[1] 杨忠[1] 黄宵宁[2] 吴怀群[1] 

机构地区：[1]南京航空航天大学自动化学院,江苏南京210016 [2]南京工程学院电力工程学院,江苏南京211167

出　　处：《应用科技》2012年第5期36-39,共4页Applied Science and Technology

基　　金：高等学校博士学科点专项科研基金资助项目(20113218110013)

摘　　要：在分析输电线路无人机巡检系统航拍图像中电力线特征的基础上,针对单一Ratio算子和Hough变换提取电力线效率低、效果差的不足,提出一种基于特征检测的电力线提取算法.首先,运用Ratio算子检测直线像素点并运用Hough变换的方法提取图像中的直线信息;然后,运用特征检测算法对电力线进行提取;最后,通过实验验证了所提算法能够从复杂的自然背景中完整提取电力线,同时能有效避免漏检、误检等情况,并有较好的噪声抑制能力.该研究为电力线典型故障的检测识别提供了有力支撑,具有较好的工程应用价值.On the basis of analysis of the characteristics of the power lines of aerial images in the unmanned aerial vehicle inspection system, a method is presented to extract power lines based on the features to overcome the shortcomings of low efficiency and poor results when using simple Ratio edge detector and hough transform to extract power lines. Firstly, the method utilizes Ratio edge detector to get edge points of lines and Hough transform to extract straight lines; secondly, the feature detection algorithm is used to extract power lines; at last, the result of the experiment indicates the algorithm can completely extract power lines from aerial images with complicated natural background and it has strong antinoise ability. The algorithm provides guarantee for detecting the typical defects and fault of the power lines, so it has a good practical value.

关 键 词：航拍图像 电力线提取 Ratio算子 HOUGH变换 特征检测 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]