Unmanned Aerial Vehicle Transmission Line Detection Based on Unsupervised Domain Adaptation Algorithm

Ting LI; Deyu AN; Jiahua LAI

doi:10.12450/j.gdzjs.202404014

Optoelectronic Technology, Volume. 44, Issue 4, 345(2024)

Unmanned Aerial Vehicle Transmission Line Detection Based on Unsupervised Domain Adaptation Algorithm

Ting LI^1,2,3, Deyu AN^2,3, and Jiahua LAI³

Author Affiliations

¹College of Computer and Cyber Security, Fujian Normal University, Fuzhou 3507, CHN

²Fujian College, University of Chinese Academy of Sciences, Quanzhou Fujian 3600, CHN

³Quanzhou Institute of Equipment Manufacturing, Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, Quanzhou Fujian 62200, CHN

show less

Abstract Get PDF(in Chinese)

Figures & Tables(6)

Fig. 1. AirSim simulated image and corresponding power line semantic annotation information

Download full size

View in Article

Fig. 2. Schematic of the mixture method

Download full size

View in Article

Fig. 3. Quantitative visualization comparison

Download full size

View in Article

Table 1. Performance of different Semantic segmentation methods on the TTPLA dataset
View table
View in Article
Table 1. Performance of different Semantic segmentation methods on the TTPLA dataset
方法骨干网络参数量/Million 交并比
Deeplab Resnet⁃101 68.8 70.48
PSPNet Resnet⁃101 68.8 70.59
U⁃Net Resnet⁃101 68.8 72.21
Segformer MiT⁃B5 84.7 75.38

Table 2. Performance comparison of cross domain Mix enhancement from AirSim data to TTPLA data （IoU）
View table
View in Article
Table 2. Performance comparison of cross domain Mix enhancement from AirSim data to TTPLA data （IoU）
网络 UDA方法 w/o CDME w/ CDME
DAFormer DAFormer 28.88 29.64
DAFormer HRDA 28.42 31.33
DAFormer MIC 30.68 32.26

Table 3. Performance comparison of style adaptive instance normalization combined with cross domain mix enhancement method from AirSim data to TTPLA dataset （IoU）
View table
View in Article
Table 3. Performance comparison of style adaptive instance normalization combined with cross domain mix enhancement method from AirSim data to TTPLA dataset （IoU）
网络 UDA方法 CDME SAIN+CDME
DAFormer DAFormer 29.64 32.11
DAFormer HRDA 31.33 34.54
DAFormer MIC 32.26 36.38

Tools

Get Citation

Copy Citation Text

Ting LI, Deyu AN, Jiahua LAI. Unmanned Aerial Vehicle Transmission Line Detection Based on Unsupervised Domain Adaptation Algorithm[J]. Optoelectronic Technology, 2024, 44(4): 345

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category:

Received: Dec. 9, 2023

Accepted: --

Published Online: Mar. 5, 2025

The Author Email:

DOI:10.12450/j.gdzjs.202404014

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology

Table 1. Performance of different Semantic segmentation methods on the TTPLA dataset

Table 1. Performance of different Semantic segmentation methods on the TTPLA dataset

Table 2. Performance comparison of cross domain Mix enhancement from AirSim data to TTPLA data （IoU）

Table 2. Performance comparison of cross domain Mix enhancement from AirSim data to TTPLA data （IoU）

Table 3. Performance comparison of style adaptive instance normalization combined with cross domain mix enhancement method from AirSim data to TTPLA dataset （IoU）

Table 3. Performance comparison of style adaptive instance normalization combined with cross domain mix enhancement method from AirSim data to TTPLA dataset （IoU）