Research on adaptive detection technology for pin defects in transmission lines

被引：0

作者：

Zhao L. ^{[1
]}

Liu C. ^{[2
]}

Zhang Z. ^{[2
]}

Qu H. ^{[2
]}

机构：

[1] School of Control and Computer Engineering, North China Electric Power University, Beijing

[2] Information College, North China University of Technology, Beijing

来源：

Huazhong Keji Daxue Xuebao (Ziran Kexue Ban)/Journal of Huazhong University of Science and Technology (Natural Science Edition) | 2023年 / 51卷 / 02期

关键词：

adaptive detection; aLRP loss; deformable convolution; pin defect; transmission line;

D O I：

10.13245/j.hust.230211

中图分类号：

学科分类号：

摘要：

In view of the detection difficulty for the target with small characteristics size in a complex background environment，an optimized algorithm based on the Faster R-CNN framework was proposed in this study．First，in the feature extraction module，the residual structure was superimposed in residual network，which could enlarge the receptive field and improve the feature extraction ability． And then the traditional convolution kernel was superseded by the deformable convolution kernel in the convolution operation，which could realize adaptive feature extraction for the target with large spatial deformation．Finally，the classification and regression tasks were balanced by average localization recall precision loss (aLRP loss)，which made an improvement on the accuracy detection of pin defect，and the optimal localization recall precision (oLRP) was used to measure the test results．Experiments of the test set shown that the mean average precision (mAP) of the optimized algorithm is 7.6% higher than that of the baseline algorithm，and the average value of oLRP of optimized algorithm is 7.1% lower than that of algorithm without localization ranking. © 2023 Huazhong University of Science and Technology. All rights reserved.

引用

页码：109 / 115+160

共 15 条

[1] TOTH J，, GILPIN-JACKSON A．, Smart view for a smart grid—unmanned aerial vehicles for transmission lines[C], Proc of the IEEE International Conference on Applied Robotics for the Power Industry, pp. 1-6, (2010)
[2] 43, 3, pp. 1057-1069, (2020)
[3] LOWE D．, Distinctive image features from scaleinvariant keypoints[J], International Journal of Computer Vision, 60, 2, pp. 91-110, (2004)
[4] BAY H，, TUYTELAARS T，, GOOL L V．, SURF：speeded up robust features[C], Proc of European Conference on Computer Vision, pp. 404-417, (2006)
[5] RUBLEE E，, RABAUD V，, KONOLIGE K, ORB：an efficient alternative to SIFT or SURF[C], Proc of the IEEE Conference on Computer Vision, pp. 2564-2571, (2011)
[6] 48, 1, pp. 37-41
[7] REDMON J, DIVVALA S, GIRSHICK R, You only look once：unified，real-time object detection [C], Proc of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 779-788, (2016)
[8] LIU W，, ANGUELOV D，, ERHAN D, SSD：single shot multibox detector[C], Proc of European Conference on Computer Vision, pp. 21-37, (2016)
[9] 48, 9, pp. 70-76, (2020)
[10] REN S，, HE K, GIRSHICK R, Faster R-CNN：towards real-time object detection with region proposal networks[J], IEEE Transactions on Pattern Analysis & Machine Intelligence, 39, 6, pp. 1137-1149, (2017)

← 1 2 →