摘要
利用无人机进行输电线路巡检是近几年国内外研究的热点技术之一,其优点是在无需拉闸断电的情况下,即可对输电线路进行检测,对其故障进行判别。根据输电线路设备的特征,应用图像处理与模式识别技术,提出了一种识别绝缘子、防震锤和输电塔的方法。该方法先采用中值滤波、膨胀和腐蚀等方法对灰度化后的航拍图像进行预处理,然后提取预处理后图像的小波特征值,最后采用AP(Affinity Propagation)聚类方法对目标图像进行分类与识别。实验结果表明,所提出的方法能够有效的识别绝缘子、防震锤和输电塔等目标,具有较好的鲁棒性和准确性。
In recent years, the use of unmanned aerial vehicle（UAV） for power transmission line detection is a hot technology, and it can detect the transmission line and discriminate the failure without power off. According to the features of power transmission line equipments, a method of equipments recognition based on image processing and pattern recognition was put forward. It preprocessed the aerial images by median filtering, dilating and eroding. Then, wavelet eigenvalues of preprocessed images were calculated. Finally, the objects, such as insulator, vibration damper and transmission tower, were classified and recognized by affinity propagation （AP） clustering algorithm. The results of experiments show that this method can recognize the insulator, vibration damper and transmission tower, and it has better robustness, accuracy and validity.
出处
《红外与激光工程》
EI
CSCD
北大核心
2013年第7期1940-1944,共5页
Infrared and Laser Engineering
基金
国家电网公司科技项目(无人飞行器巡检技术研究)
关键词
目标识别
小波特征
AP聚类
无人机
object recognition
wavelet feature
AP clustering
UAV
