Automatic image detection of multi-type surface defects on wind turbine blades based on cascade deep learning network
A safe operation protocol of the wind blades is a critical factor to ensure the stability of a wind turbine. Sensors are most commonly applied for defect detection on wind turbine blades (WTBs). However, due to the high cost and the sensitivity to stochastic noise, computer vision-guided automatic d...
Gespeichert in:
Veröffentlicht in: | Intelligent data analysis 2021-03, Vol.25 (2), p.463-482 |
---|---|
Hauptverfasser: | , , , |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
container_end_page | 482 |
---|---|
container_issue | 2 |
container_start_page | 463 |
container_title | Intelligent data analysis |
container_volume | 25 |
creator | Mao, Yulin Wang, Shuangxin Yu, Dingli Zhao, Juchao |
description | A safe operation protocol of the wind blades is a critical factor to ensure the stability of a wind turbine. Sensors are most commonly applied for defect detection on wind turbine blades (WTBs). However, due to the high cost and the sensitivity to stochastic noise, computer vision-guided automatic detection remains a challenge for surface defect detection on WTBs in particularly, its accuracy in locating defects is yet to be optimized. In this paper, we developed a visual inspection model that can automatically and precisely classify and locate the surface defects, through the utilization of a deep learning framework based on the Cascade R-CNN. In order to obtain high mean average precision (mAP) according to the characteristics of the dataset, a model named Contextual Aligned-Deformable Cascade R-CNN (CAD Cascade R-CNN) using improved strategies of transfer learning, Deformable Convolution and Deformable RoI Align, as well as context information fusion is proposed and a dataset with surface defects categorized and labeled as crack, breakage and oil pollution is generated. Moreover to alleviate the problem of false detection under a complex background, an improved bisecting k-means is presented during the test process. The adaptability and generalization of the proposed CAD Cascade R-CNN model were validated by each type of defects in dataset and different IoU thresholds, whereas, each of the above improved strategies was verified by gradual ablation experiments. Finally experiments that compared with the baseline Cascade R-CNN, Faster R-CNN and YOLO-v3 demonstrate its superiority over these existing approaches with a maximum of 92.1% mAP. |
doi_str_mv | 10.3233/IDA-205143 |
format | Article |
fullrecord | <record><control><sourceid>proquest_cross</sourceid><recordid>TN_cdi_proquest_journals_2499050193</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2499050193</sourcerecordid><originalsourceid>FETCH-LOGICAL-c222t-67be4c903decc1a23fdd47bf57496d0c85189a2eba8cf33823e2eb6238643b83</originalsourceid><addsrcrecordid>eNotkMtKAzEUhoMoWKsbnyDgTojmNjOZZalWCwU3XbgbMslJmTqdGXOh9O1Nqatz-T_OgQ-hR0ZfBBfidf22IJwWTIorNGNFxYhkXF3nnipFZFl936K7EPaUUsmpnKG0SHE86NgZ3B30DrCFCCZ244BHhw-pjx2JpwlwSN5pc85dzgPOwLEbLI7Jt90AuO21hYBbHcCeQ6ODyZvMw4R70H7ohh0eIB5H_3OPbpzuAzz81znart63y0-y-fpYLxcbYjjnkZRVC9LUVFgwhmkunLWyal1Rybq01KiCqVpzaLUyTgjFBeSh5EKVUrRKzNHT5ezkx98EITb7Mfkhf2y4rGtaUFaLTD1fKOPHEDy4ZvLZhT81jDZnq0222lysij-gLGtQ</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2499050193</pqid></control><display><type>article</type><title>Automatic image detection of multi-type surface defects on wind turbine blades based on cascade deep learning network</title><source>Business Source Complete</source><creator>Mao, Yulin ; Wang, Shuangxin ; Yu, Dingli ; Zhao, Juchao</creator><creatorcontrib>Mao, Yulin ; Wang, Shuangxin ; Yu, Dingli ; Zhao, Juchao</creatorcontrib><description>A safe operation protocol of the wind blades is a critical factor to ensure the stability of a wind turbine. Sensors are most commonly applied for defect detection on wind turbine blades (WTBs). However, due to the high cost and the sensitivity to stochastic noise, computer vision-guided automatic detection remains a challenge for surface defect detection on WTBs in particularly, its accuracy in locating defects is yet to be optimized. In this paper, we developed a visual inspection model that can automatically and precisely classify and locate the surface defects, through the utilization of a deep learning framework based on the Cascade R-CNN. In order to obtain high mean average precision (mAP) according to the characteristics of the dataset, a model named Contextual Aligned-Deformable Cascade R-CNN (CAD Cascade R-CNN) using improved strategies of transfer learning, Deformable Convolution and Deformable RoI Align, as well as context information fusion is proposed and a dataset with surface defects categorized and labeled as crack, breakage and oil pollution is generated. Moreover to alleviate the problem of false detection under a complex background, an improved bisecting k-means is presented during the test process. The adaptability and generalization of the proposed CAD Cascade R-CNN model were validated by each type of defects in dataset and different IoU thresholds, whereas, each of the above improved strategies was verified by gradual ablation experiments. Finally experiments that compared with the baseline Cascade R-CNN, Faster R-CNN and YOLO-v3 demonstrate its superiority over these existing approaches with a maximum of 92.1% mAP.</description><identifier>ISSN: 1088-467X</identifier><identifier>EISSN: 1571-4128</identifier><identifier>DOI: 10.3233/IDA-205143</identifier><language>eng</language><publisher>Amsterdam: IOS Press BV</publisher><subject>Ablation ; Breakage ; Computer vision ; Convolution ; Data integration ; Datasets ; Deep learning ; Deformation ; Formability ; Image detection ; Inspection ; Noise sensitivity ; Oil pollution ; Surface defects ; Turbine blades ; Wind turbines</subject><ispartof>Intelligent data analysis, 2021-03, Vol.25 (2), p.463-482</ispartof><rights>Copyright IOS Press BV 2021</rights><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c222t-67be4c903decc1a23fdd47bf57496d0c85189a2eba8cf33823e2eb6238643b83</citedby><cites>FETCH-LOGICAL-c222t-67be4c903decc1a23fdd47bf57496d0c85189a2eba8cf33823e2eb6238643b83</cites></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>314,776,780,27903,27904</link.rule.ids></links><search><creatorcontrib>Mao, Yulin</creatorcontrib><creatorcontrib>Wang, Shuangxin</creatorcontrib><creatorcontrib>Yu, Dingli</creatorcontrib><creatorcontrib>Zhao, Juchao</creatorcontrib><title>Automatic image detection of multi-type surface defects on wind turbine blades based on cascade deep learning network</title><title>Intelligent data analysis</title><description>A safe operation protocol of the wind blades is a critical factor to ensure the stability of a wind turbine. Sensors are most commonly applied for defect detection on wind turbine blades (WTBs). However, due to the high cost and the sensitivity to stochastic noise, computer vision-guided automatic detection remains a challenge for surface defect detection on WTBs in particularly, its accuracy in locating defects is yet to be optimized. In this paper, we developed a visual inspection model that can automatically and precisely classify and locate the surface defects, through the utilization of a deep learning framework based on the Cascade R-CNN. In order to obtain high mean average precision (mAP) according to the characteristics of the dataset, a model named Contextual Aligned-Deformable Cascade R-CNN (CAD Cascade R-CNN) using improved strategies of transfer learning, Deformable Convolution and Deformable RoI Align, as well as context information fusion is proposed and a dataset with surface defects categorized and labeled as crack, breakage and oil pollution is generated. Moreover to alleviate the problem of false detection under a complex background, an improved bisecting k-means is presented during the test process. The adaptability and generalization of the proposed CAD Cascade R-CNN model were validated by each type of defects in dataset and different IoU thresholds, whereas, each of the above improved strategies was verified by gradual ablation experiments. Finally experiments that compared with the baseline Cascade R-CNN, Faster R-CNN and YOLO-v3 demonstrate its superiority over these existing approaches with a maximum of 92.1% mAP.</description><subject>Ablation</subject><subject>Breakage</subject><subject>Computer vision</subject><subject>Convolution</subject><subject>Data integration</subject><subject>Datasets</subject><subject>Deep learning</subject><subject>Deformation</subject><subject>Formability</subject><subject>Image detection</subject><subject>Inspection</subject><subject>Noise sensitivity</subject><subject>Oil pollution</subject><subject>Surface defects</subject><subject>Turbine blades</subject><subject>Wind turbines</subject><issn>1088-467X</issn><issn>1571-4128</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2021</creationdate><recordtype>article</recordtype><recordid>eNotkMtKAzEUhoMoWKsbnyDgTojmNjOZZalWCwU3XbgbMslJmTqdGXOh9O1Nqatz-T_OgQ-hR0ZfBBfidf22IJwWTIorNGNFxYhkXF3nnipFZFl936K7EPaUUsmpnKG0SHE86NgZ3B30DrCFCCZ244BHhw-pjx2JpwlwSN5pc85dzgPOwLEbLI7Jt90AuO21hYBbHcCeQ6ODyZvMw4R70H7ohh0eIB5H_3OPbpzuAzz81znart63y0-y-fpYLxcbYjjnkZRVC9LUVFgwhmkunLWyal1Rybq01KiCqVpzaLUyTgjFBeSh5EKVUrRKzNHT5ezkx98EITb7Mfkhf2y4rGtaUFaLTD1fKOPHEDy4ZvLZhT81jDZnq0222lysij-gLGtQ</recordid><startdate>20210304</startdate><enddate>20210304</enddate><creator>Mao, Yulin</creator><creator>Wang, Shuangxin</creator><creator>Yu, Dingli</creator><creator>Zhao, Juchao</creator><general>IOS Press BV</general><scope>AAYXX</scope><scope>CITATION</scope><scope>7SC</scope><scope>8FD</scope><scope>JQ2</scope><scope>L7M</scope><scope>L~C</scope><scope>L~D</scope></search><sort><creationdate>20210304</creationdate><title>Automatic image detection of multi-type surface defects on wind turbine blades based on cascade deep learning network</title><author>Mao, Yulin ; Wang, Shuangxin ; Yu, Dingli ; Zhao, Juchao</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c222t-67be4c903decc1a23fdd47bf57496d0c85189a2eba8cf33823e2eb6238643b83</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2021</creationdate><topic>Ablation</topic><topic>Breakage</topic><topic>Computer vision</topic><topic>Convolution</topic><topic>Data integration</topic><topic>Datasets</topic><topic>Deep learning</topic><topic>Deformation</topic><topic>Formability</topic><topic>Image detection</topic><topic>Inspection</topic><topic>Noise sensitivity</topic><topic>Oil pollution</topic><topic>Surface defects</topic><topic>Turbine blades</topic><topic>Wind turbines</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Mao, Yulin</creatorcontrib><creatorcontrib>Wang, Shuangxin</creatorcontrib><creatorcontrib>Yu, Dingli</creatorcontrib><creatorcontrib>Zhao, Juchao</creatorcontrib><collection>CrossRef</collection><collection>Computer and Information Systems Abstracts</collection><collection>Technology Research Database</collection><collection>ProQuest Computer Science Collection</collection><collection>Advanced Technologies Database with Aerospace</collection><collection>Computer and Information Systems Abstracts Academic</collection><collection>Computer and Information Systems Abstracts Professional</collection><jtitle>Intelligent data analysis</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Mao, Yulin</au><au>Wang, Shuangxin</au><au>Yu, Dingli</au><au>Zhao, Juchao</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Automatic image detection of multi-type surface defects on wind turbine blades based on cascade deep learning network</atitle><jtitle>Intelligent data analysis</jtitle><date>2021-03-04</date><risdate>2021</risdate><volume>25</volume><issue>2</issue><spage>463</spage><epage>482</epage><pages>463-482</pages><issn>1088-467X</issn><eissn>1571-4128</eissn><abstract>A safe operation protocol of the wind blades is a critical factor to ensure the stability of a wind turbine. Sensors are most commonly applied for defect detection on wind turbine blades (WTBs). However, due to the high cost and the sensitivity to stochastic noise, computer vision-guided automatic detection remains a challenge for surface defect detection on WTBs in particularly, its accuracy in locating defects is yet to be optimized. In this paper, we developed a visual inspection model that can automatically and precisely classify and locate the surface defects, through the utilization of a deep learning framework based on the Cascade R-CNN. In order to obtain high mean average precision (mAP) according to the characteristics of the dataset, a model named Contextual Aligned-Deformable Cascade R-CNN (CAD Cascade R-CNN) using improved strategies of transfer learning, Deformable Convolution and Deformable RoI Align, as well as context information fusion is proposed and a dataset with surface defects categorized and labeled as crack, breakage and oil pollution is generated. Moreover to alleviate the problem of false detection under a complex background, an improved bisecting k-means is presented during the test process. The adaptability and generalization of the proposed CAD Cascade R-CNN model were validated by each type of defects in dataset and different IoU thresholds, whereas, each of the above improved strategies was verified by gradual ablation experiments. Finally experiments that compared with the baseline Cascade R-CNN, Faster R-CNN and YOLO-v3 demonstrate its superiority over these existing approaches with a maximum of 92.1% mAP.</abstract><cop>Amsterdam</cop><pub>IOS Press BV</pub><doi>10.3233/IDA-205143</doi><tpages>20</tpages></addata></record> |
fulltext | fulltext |
identifier | ISSN: 1088-467X |
ispartof | Intelligent data analysis, 2021-03, Vol.25 (2), p.463-482 |
issn | 1088-467X 1571-4128 |
language | eng |
recordid | cdi_proquest_journals_2499050193 |
source | Business Source Complete |
subjects | Ablation Breakage Computer vision Convolution Data integration Datasets Deep learning Deformation Formability Image detection Inspection Noise sensitivity Oil pollution Surface defects Turbine blades Wind turbines |
title | Automatic image detection of multi-type surface defects on wind turbine blades based on cascade deep learning network |
url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-27T13%3A41%3A57IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_cross&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Automatic%20image%20detection%20of%20multi-type%20surface%20defects%20on%20wind%20turbine%20blades%20based%20on%20cascade%20deep%20learning%20network&rft.jtitle=Intelligent%20data%20analysis&rft.au=Mao,%20Yulin&rft.date=2021-03-04&rft.volume=25&rft.issue=2&rft.spage=463&rft.epage=482&rft.pages=463-482&rft.issn=1088-467X&rft.eissn=1571-4128&rft_id=info:doi/10.3233/IDA-205143&rft_dat=%3Cproquest_cross%3E2499050193%3C/proquest_cross%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2499050193&rft_id=info:pmid/&rfr_iscdi=true |