Cover Image

AUTONOMOUS POWER LINE INSPECTION USING COMPUTER VISION

This project aims to construct an autonomous power line inspection system using computer vision to classify and localise the normal and abnormal insulators. The traditional inspection methods, for instance, foot patrol inspection and helicopter-assisted inspection are time-consuming and dangerous. D...

Full description

Saved in:
Bibliographic Details
Main Author: LAW, JIN MING
Format: Final Year Project Report
Language:English
English
Published: Universiti Malaysia Sarawak, (UNIMAS) 2022
Subjects:
TA Engineering (General). Civil engineering (General)
Online Access:http://ir.unimas.my/id/eprint/40114/1/Law%20Jin%20Ming%2024pgs.pdf
http://ir.unimas.my/id/eprint/40114/5/Law%20Jin%20Ming%20ft.pdf
http://ir.unimas.my/id/eprint/40114/
Tags: Add Tag
No Tags, Be the first to tag this record!
id my.unimas.ir.40114
record_format eprints
spelling my.unimas.ir.401142024-01-10T06:58:24Z http://ir.unimas.my/id/eprint/40114/ AUTONOMOUS POWER LINE INSPECTION USING COMPUTER VISION LAW, JIN MING TA Engineering (General). Civil engineering (General) This project aims to construct an autonomous power line inspection system using computer vision to classify and localise the normal and abnormal insulators. The traditional inspection methods, for instance, foot patrol inspection and helicopter-assisted inspection are time-consuming and dangerous. DenseNet-201 model is proposed as the base network to perform insulator fault detection autonomously. An algorithm with DenseNet-201 backbone consisting of two branches which are class label classification and bounding box regression is developed. The developed algorithm is trained on the augmented Chinese Power Line Insulator Dataset (CPLID) that consisted of normal and missing cap insulator images. The prediction results in classification accuracy of 100%. The average precision of detecting normal insulator and insulator missing cap has higher performance at threshold 0.3 which are 100% and 66.61%. The mean average precision at thresholds 0.3 and 0.5 are 83.31% and 64.62% respectively. The experimental results on the augmented CPLID dataset denote that the proposed model has high classification accuracy and it outperforms the ResNet model. Universiti Malaysia Sarawak, (UNIMAS) 2022 Final Year Project Report NonPeerReviewed text en http://ir.unimas.my/id/eprint/40114/1/Law%20Jin%20Ming%2024pgs.pdf text en http://ir.unimas.my/id/eprint/40114/5/Law%20Jin%20Ming%20ft.pdf LAW, JIN MING (2022) AUTONOMOUS POWER LINE INSPECTION USING COMPUTER VISION. [Final Year Project Report] (Unpublished)
institution Universiti Malaysia Sarawak
building Centre for Academic Information Services (CAIS)
collection Institutional Repository
continent Asia
country Malaysia
content_provider Universiti Malaysia Sarawak
content_source UNIMAS Institutional Repository
url_provider http://ir.unimas.my/
language English
English
topic TA Engineering (General). Civil engineering (General)
spellingShingle TA Engineering (General). Civil engineering (General)
LAW, JIN MING
AUTONOMOUS POWER LINE INSPECTION USING COMPUTER VISION
description This project aims to construct an autonomous power line inspection system using computer vision to classify and localise the normal and abnormal insulators. The traditional inspection methods, for instance, foot patrol inspection and helicopter-assisted inspection are time-consuming and dangerous. DenseNet-201 model is proposed as the base network to perform insulator fault detection autonomously. An algorithm with DenseNet-201 backbone consisting of two branches which are class label classification and bounding box regression is developed. The developed algorithm is trained on the augmented Chinese Power Line Insulator Dataset (CPLID) that consisted of normal and missing cap insulator images. The prediction results in classification accuracy of 100%. The average precision of detecting normal insulator and insulator missing cap has higher performance at threshold 0.3 which are 100% and 66.61%. The mean average precision at thresholds 0.3 and 0.5 are 83.31% and 64.62% respectively. The experimental results on the augmented CPLID dataset denote that the proposed model has high classification accuracy and it outperforms the ResNet model.
format Final Year Project Report
author LAW, JIN MING
author_facet LAW, JIN MING
author_sort LAW, JIN MING
title AUTONOMOUS POWER LINE INSPECTION USING COMPUTER VISION
title_short AUTONOMOUS POWER LINE INSPECTION USING COMPUTER VISION
title_full AUTONOMOUS POWER LINE INSPECTION USING COMPUTER VISION
title_fullStr AUTONOMOUS POWER LINE INSPECTION USING COMPUTER VISION
title_full_unstemmed AUTONOMOUS POWER LINE INSPECTION USING COMPUTER VISION
title_sort autonomous power line inspection using computer vision
publisher Universiti Malaysia Sarawak, (UNIMAS)
publishDate 2022
url http://ir.unimas.my/id/eprint/40114/1/Law%20Jin%20Ming%2024pgs.pdf
http://ir.unimas.my/id/eprint/40114/5/Law%20Jin%20Ming%20ft.pdf
http://ir.unimas.my/id/eprint/40114/
_version_ 1789430346069049344
score 13.211869

Similar Items