The current methods for detecting foreign objects on transmission lines suffer from low effi-ciency and accuracy,a three-step approach for transmission line extraction and a foreign object detection algorithm based on rectangular region expansion are proposed.Firstly,a variety of feature extraction opera-tors are fused to detect the edge features of the image,and Gaussian pyramid is used for feature optimiza-tion to accelerate the speed of feature extraction;secondly,the transmission lines are extracted based on the similarity mechanism of linear feature recognition,based on the extraction of transmission lines,the detec-tion and location of foreign objects on transmission lines are realized by means of rectangular region expan-sion detection;finally,it is proved through experiments that the detection accuracy of the method proposed in this paper can reach 99.21％and 93.89％for large and small target foreign objects respectively,and its corresponding detection time reaches 120 ms and 60 ms,respectively.The results show that compared with the existing conventional methods,the proposed method has a higher accuracy in the detection of power lines and foreign objects,and can greatly improve the safety and efficiency of foreign objects detection.
关键词
特征提取/异物检测/直线机制/区域膨胀/输电线路
Key words
feature extraction/foreign objects detection/linear mechanism/regional expansion/trans-mission line
维普
万方数据