Power Distribution Insulators Classification Using Image Hybrid Deep Learning. Filho, E. F. S., Prates, R. M., Ramos, R. P., & Cardoso, J. S. In 2019 27th European Signal Processing Conference (EUSIPCO), pages 1-5, Sep., 2019. Paper doi abstract bibtex The Overhead Power Distribution Lines present a wide range of insulator components, which have different shapes and types of building materials. These components are usually exposed to weather and operational conditions that may cause deviations in their shapes, colors or textures. These changes might hinder the development of automatic systems for visual inspection. In this perspective, this work presents a robust methodology for image classification, which aims at the efficient distribution insulator class identification, regardless of its degradation level. This work can be characterized by the following steps: implementation of Convolutional Neural Network (CNN); transfer learning; attribute vector acquisition and design of hybrid classifier architectures to improve the discrimination efficiency. In summary, a previously trained CNN goes through a fine tuning stage for later use as a feature extractor for training a new set of classifiers. A comparative study was conducted to identify which classifier architecture obtained the best discrimination performance for non-conforming components. The proposed methodology showed a significant improvement in classification performance, obtaining 95% overall accuracy in the identification of non-conforming component classes.
@InProceedings{8903139,
author = {E. F. S. Filho and R. M. Prates and R. P. Ramos and J. S. Cardoso},
booktitle = {2019 27th European Signal Processing Conference (EUSIPCO)},
title = {Power Distribution Insulators Classification Using Image Hybrid Deep Learning},
year = {2019},
pages = {1-5},
abstract = {The Overhead Power Distribution Lines present a wide range of insulator components, which have different shapes and types of building materials. These components are usually exposed to weather and operational conditions that may cause deviations in their shapes, colors or textures. These changes might hinder the development of automatic systems for visual inspection. In this perspective, this work presents a robust methodology for image classification, which aims at the efficient distribution insulator class identification, regardless of its degradation level. This work can be characterized by the following steps: implementation of Convolutional Neural Network (CNN); transfer learning; attribute vector acquisition and design of hybrid classifier architectures to improve the discrimination efficiency. In summary, a previously trained CNN goes through a fine tuning stage for later use as a feature extractor for training a new set of classifiers. A comparative study was conducted to identify which classifier architecture obtained the best discrimination performance for non-conforming components. The proposed methodology showed a significant improvement in classification performance, obtaining 95% overall accuracy in the identification of non-conforming component classes.},
keywords = {convolutional neural nets;feature extraction;image classification;inspection;insulators;learning (artificial intelligence);power distribution lines;power engineering computing;power overhead lines;fine tuning stage;nonconforming components;image hybrid deep learning;overhead power distribution lines;insulator components;building materials;operational conditions;automatic systems;visual inspection;image classification performance;degradation level;convolutional neural network;transfer learning;attribute vector acquisition;hybrid classifier architectures;power distribution insulators classification;distribution insulator class identification;feature extractor;Insulators;Training;Vegetation;Inspection;Support vector machines;Visualization;Convolutional neural networks;distribution insulators;convolutional neural network (CNN);transfer learning;hybrid classifiers},
doi = {10.23919/EUSIPCO.2019.8903139},
issn = {2076-1465},
month = {Sep.},
url = {https://www.eurasip.org/proceedings/eusipco/eusipco2019/proceedings/papers/1570533893.pdf},
}
Downloads: 0
{"_id":"k9TJKiXBKJD533d6g","bibbaseid":"filho-prates-ramos-cardoso-powerdistributioninsulatorsclassificationusingimagehybriddeeplearning-2019","authorIDs":[],"author_short":["Filho, E. F. S.","Prates, R. M.","Ramos, R. P.","Cardoso, J. S."],"bibdata":{"bibtype":"inproceedings","type":"inproceedings","author":[{"firstnames":["E.","F.","S."],"propositions":[],"lastnames":["Filho"],"suffixes":[]},{"firstnames":["R.","M."],"propositions":[],"lastnames":["Prates"],"suffixes":[]},{"firstnames":["R.","P."],"propositions":[],"lastnames":["Ramos"],"suffixes":[]},{"firstnames":["J.","S."],"propositions":[],"lastnames":["Cardoso"],"suffixes":[]}],"booktitle":"2019 27th European Signal Processing Conference (EUSIPCO)","title":"Power Distribution Insulators Classification Using Image Hybrid Deep Learning","year":"2019","pages":"1-5","abstract":"The Overhead Power Distribution Lines present a wide range of insulator components, which have different shapes and types of building materials. These components are usually exposed to weather and operational conditions that may cause deviations in their shapes, colors or textures. These changes might hinder the development of automatic systems for visual inspection. In this perspective, this work presents a robust methodology for image classification, which aims at the efficient distribution insulator class identification, regardless of its degradation level. This work can be characterized by the following steps: implementation of Convolutional Neural Network (CNN); transfer learning; attribute vector acquisition and design of hybrid classifier architectures to improve the discrimination efficiency. In summary, a previously trained CNN goes through a fine tuning stage for later use as a feature extractor for training a new set of classifiers. A comparative study was conducted to identify which classifier architecture obtained the best discrimination performance for non-conforming components. The proposed methodology showed a significant improvement in classification performance, obtaining 95% overall accuracy in the identification of non-conforming component classes.","keywords":"convolutional neural nets;feature extraction;image classification;inspection;insulators;learning (artificial intelligence);power distribution lines;power engineering computing;power overhead lines;fine tuning stage;nonconforming components;image hybrid deep learning;overhead power distribution lines;insulator components;building materials;operational conditions;automatic systems;visual inspection;image classification performance;degradation level;convolutional neural network;transfer learning;attribute vector acquisition;hybrid classifier architectures;power distribution insulators classification;distribution insulator class identification;feature extractor;Insulators;Training;Vegetation;Inspection;Support vector machines;Visualization;Convolutional neural networks;distribution insulators;convolutional neural network (CNN);transfer learning;hybrid classifiers","doi":"10.23919/EUSIPCO.2019.8903139","issn":"2076-1465","month":"Sep.","url":"https://www.eurasip.org/proceedings/eusipco/eusipco2019/proceedings/papers/1570533893.pdf","bibtex":"@InProceedings{8903139,\n author = {E. F. S. Filho and R. M. Prates and R. P. Ramos and J. S. Cardoso},\n booktitle = {2019 27th European Signal Processing Conference (EUSIPCO)},\n title = {Power Distribution Insulators Classification Using Image Hybrid Deep Learning},\n year = {2019},\n pages = {1-5},\n abstract = {The Overhead Power Distribution Lines present a wide range of insulator components, which have different shapes and types of building materials. These components are usually exposed to weather and operational conditions that may cause deviations in their shapes, colors or textures. These changes might hinder the development of automatic systems for visual inspection. In this perspective, this work presents a robust methodology for image classification, which aims at the efficient distribution insulator class identification, regardless of its degradation level. This work can be characterized by the following steps: implementation of Convolutional Neural Network (CNN); transfer learning; attribute vector acquisition and design of hybrid classifier architectures to improve the discrimination efficiency. In summary, a previously trained CNN goes through a fine tuning stage for later use as a feature extractor for training a new set of classifiers. A comparative study was conducted to identify which classifier architecture obtained the best discrimination performance for non-conforming components. The proposed methodology showed a significant improvement in classification performance, obtaining 95% overall accuracy in the identification of non-conforming component classes.},\n keywords = {convolutional neural nets;feature extraction;image classification;inspection;insulators;learning (artificial intelligence);power distribution lines;power engineering computing;power overhead lines;fine tuning stage;nonconforming components;image hybrid deep learning;overhead power distribution lines;insulator components;building materials;operational conditions;automatic systems;visual inspection;image classification performance;degradation level;convolutional neural network;transfer learning;attribute vector acquisition;hybrid classifier architectures;power distribution insulators classification;distribution insulator class identification;feature extractor;Insulators;Training;Vegetation;Inspection;Support vector machines;Visualization;Convolutional neural networks;distribution insulators;convolutional neural network (CNN);transfer learning;hybrid classifiers},\n doi = {10.23919/EUSIPCO.2019.8903139},\n issn = {2076-1465},\n month = {Sep.},\n url = {https://www.eurasip.org/proceedings/eusipco/eusipco2019/proceedings/papers/1570533893.pdf},\n}\n\n","author_short":["Filho, E. F. S.","Prates, R. M.","Ramos, R. P.","Cardoso, J. S."],"key":"8903139","id":"8903139","bibbaseid":"filho-prates-ramos-cardoso-powerdistributioninsulatorsclassificationusingimagehybriddeeplearning-2019","role":"author","urls":{"Paper":"https://www.eurasip.org/proceedings/eusipco/eusipco2019/proceedings/papers/1570533893.pdf"},"keyword":["convolutional neural nets;feature extraction;image classification;inspection;insulators;learning (artificial intelligence);power distribution lines;power engineering computing;power overhead lines;fine tuning stage;nonconforming components;image hybrid deep learning;overhead power distribution lines;insulator components;building materials;operational conditions;automatic systems;visual inspection;image classification performance;degradation level;convolutional neural network;transfer learning;attribute vector acquisition;hybrid classifier architectures;power distribution insulators classification;distribution insulator class identification;feature extractor;Insulators;Training;Vegetation;Inspection;Support vector machines;Visualization;Convolutional neural networks;distribution insulators;convolutional neural network (CNN);transfer learning;hybrid classifiers"],"metadata":{"authorlinks":{}},"downloads":0},"bibtype":"inproceedings","biburl":"https://raw.githubusercontent.com/Roznn/EUSIPCO/main/eusipco2019url.bib","creationDate":"2021-02-11T19:15:22.164Z","downloads":0,"keywords":["convolutional neural nets;feature extraction;image classification;inspection;insulators;learning (artificial intelligence);power distribution lines;power engineering computing;power overhead lines;fine tuning stage;nonconforming components;image hybrid deep learning;overhead power distribution lines;insulator components;building materials;operational conditions;automatic systems;visual inspection;image classification performance;degradation level;convolutional neural network;transfer learning;attribute vector acquisition;hybrid classifier architectures;power distribution insulators classification;distribution insulator class identification;feature extractor;insulators;training;vegetation;inspection;support vector machines;visualization;convolutional neural networks;distribution insulators;convolutional neural network (cnn);transfer learning;hybrid classifiers"],"search_terms":["power","distribution","insulators","classification","using","image","hybrid","deep","learning","filho","prates","ramos","cardoso"],"title":"Power Distribution Insulators Classification Using Image Hybrid Deep Learning","year":2019,"dataSources":["NqWTiMfRR56v86wRs","r6oz3cMyC99QfiuHW"]}