Open AccessIdentification of hidden dangers in transmission line corridors based on hybrid algorithms
Author(s) -
Wei Zheng,
Chengqi Li,
Bo Yang,
Xiaobin Sun
Publication year - 2019
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1346/1/012042
Subject(s) - computer science , identification (biology) , artificial intelligence , algorithm , transmission (telecommunications) , support vector machine , line (geometry) , sample (material) , machine learning , computer vision , telecommunications , mathematics , chemistry , botany , geometry , chromatography , biology
Computers automatically recognize hidden dangers in photos taken from transmission line corridors to greatly reduce the workload of transmission line patrol personnel. However, transmission line corridors are complex and hidden dangers in the scene are diverse, which bring great difficulties to the identification of hidden dangers. To solve this problem, this paper proposes various algorithms to identify various hidden dangers separately based on their sample size and characteristics of hidden dangers. Firstly, we identify construction machinery based on CNN algorithm by making use of massive samples. Secondly, we identify fires by using features including color and texture based on SVM. Finally, we identify foreign objects on transmission lines with Bresenham algorithm based on the geometric characteristics of transmission lines. In order to verify the practicality of the above algorithms, we collect tens of thousands of photos taken by cameras deployed on transmission towers and select different samples from them to test the above algorithms separately, whose results show that our algorithms achieved acceptable recognition accuracy.