Defect detection in vehicle mirror nonplanar surfaces with multi-scale atrous single-shot detect mechanism
Surface quality inspection is important for vehicle rearview mirrors. Surface quality defects, such as bubbles, particles, cracks, and scratches, may appear during the production process. Traditionally, manual inspection is time-consuming, laborious, low in accuracy, and costly. In recent years, obj...
Gespeichert in:
Veröffentlicht in: | AIP advances 2021-07, Vol.11 (7), p.075202-075202-11 |
---|---|
Hauptverfasser: | , , , , |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
container_end_page | 075202-11 |
---|---|
container_issue | 7 |
container_start_page | 075202 |
container_title | AIP advances |
container_volume | 11 |
creator | Li, Chuanchang Cui, Guohua Zhang, Weiwei Chen, Saixuan Yang, Yinyi |
description | Surface quality inspection is important for vehicle rearview mirrors. Surface quality defects, such as bubbles, particles, cracks, and scratches, may appear during the production process. Traditionally, manual inspection is time-consuming, laborious, low in accuracy, and costly. In recent years, object detectors based on deep learning have been used for defect detection [such as single shot detectors (SSDs)], and object detectors mostly detect the details of small objects at a shallow level. However, the lack of shallow semantic information will lead to inaccuracy. The deep layer has more semantic information, but the deep layer cannot be detected due to the effect of the complete convolutional layer in the convolutional neural network. In this article, we propose an enhanced SSD method to detect micro-defects on the nonplanar surface of vehicle rearview mirrors. We call it the multi-scale atrous single-shot detector (MSASSD). Specifically, we first replace the maximum pool depth layer with an unconscious convolutional layer to expand the receiving field without reducing the size of the input image. Then, we link the shallow layer to the deep layer through the fusion block to form new and rich fusion features for object detection. Finally, we use multi-scale features (including deep features and fusion features) to predict defects. The results show that our MSASSD method can improve the average accuracy of defect detection (about 1.2% compared with the SSD method), while the detection speed is equivalent (low about two frames per second compared to the SSD method). |
doi_str_mv | 10.1063/5.0053851 |
format | Article |
fullrecord | <record><control><sourceid>proquest_doaj_</sourceid><recordid>TN_cdi_proquest_journals_2547149930</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><doaj_id>oai_doaj_org_article_a02bbd07bf344e8bbab693268fc02293</doaj_id><sourcerecordid>2547149930</sourcerecordid><originalsourceid>FETCH-LOGICAL-c428t-c673b64a4db1c079ada351f4b6476590778f4e80b45fed39067003ae56fabfe43</originalsourceid><addsrcrecordid>eNp9kUtr3DAQgE1poSHJof9A0FMLTkZP28eSNG0g0Et7FiNZymqxra0kJ-TfR9tN00CguowYPr55Nc0HCmcUFD-XZwCS95K-aY4YlX3LGVNvX_zfN6c5b6E-MVDoxVGzvXTe2UJGV2oIcSFhIXduE-zkyBxSiokscdlNuGAieU0ercvkPpQNmdephDZbrCiWFNdMclhuJ9fmTfyrJLOzG1xCnk-adx6n7E6f4nHz6-rrz4vv7c2Pb9cXX25aK1hfWqs6bpRAMRpqoRtwRC6pFzXXKTlA1_VeuB6MkN6NfADVAXB0Unk03gl-3FwfvGPErd6lMGN60BGD_pOI6VZjKvsBNQIzZoTOeC6q0xg0auBM9d4CYwOvro8H1y7F36vLRW_jmpbavmZSdFQMA4dKfTpQNsWck_PPVSno_WW01E-XqeznA5ttKLjf-DN8F9M_UO9G_z_4tfkRmD-c1g</addsrcrecordid><sourcetype>Open Website</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2547149930</pqid></control><display><type>article</type><title>Defect detection in vehicle mirror nonplanar surfaces with multi-scale atrous single-shot detect mechanism</title><source>DOAJ Directory of Open Access Journals</source><source>Elektronische Zeitschriftenbibliothek - Frei zugängliche E-Journals</source><source>Alma/SFX Local Collection</source><source>Free Full-Text Journals in Chemistry</source><creator>Li, Chuanchang ; Cui, Guohua ; Zhang, Weiwei ; Chen, Saixuan ; Yang, Yinyi</creator><creatorcontrib>Li, Chuanchang ; Cui, Guohua ; Zhang, Weiwei ; Chen, Saixuan ; Yang, Yinyi</creatorcontrib><description>Surface quality inspection is important for vehicle rearview mirrors. Surface quality defects, such as bubbles, particles, cracks, and scratches, may appear during the production process. Traditionally, manual inspection is time-consuming, laborious, low in accuracy, and costly. In recent years, object detectors based on deep learning have been used for defect detection [such as single shot detectors (SSDs)], and object detectors mostly detect the details of small objects at a shallow level. However, the lack of shallow semantic information will lead to inaccuracy. The deep layer has more semantic information, but the deep layer cannot be detected due to the effect of the complete convolutional layer in the convolutional neural network. In this article, we propose an enhanced SSD method to detect micro-defects on the nonplanar surface of vehicle rearview mirrors. We call it the multi-scale atrous single-shot detector (MSASSD). Specifically, we first replace the maximum pool depth layer with an unconscious convolutional layer to expand the receiving field without reducing the size of the input image. Then, we link the shallow layer to the deep layer through the fusion block to form new and rich fusion features for object detection. Finally, we use multi-scale features (including deep features and fusion features) to predict defects. The results show that our MSASSD method can improve the average accuracy of defect detection (about 1.2% compared with the SSD method), while the detection speed is equivalent (low about two frames per second compared to the SSD method).</description><identifier>ISSN: 2158-3226</identifier><identifier>EISSN: 2158-3226</identifier><identifier>DOI: 10.1063/5.0053851</identifier><identifier>CODEN: AAIDBI</identifier><language>eng</language><publisher>Melville: American Institute of Physics</publisher><subject>Artificial neural networks ; Defects ; Detectors ; Flaw detection ; Frames per second ; Inspection ; Object recognition ; Semantics ; Sensors ; Surface properties</subject><ispartof>AIP advances, 2021-07, Vol.11 (7), p.075202-075202-11</ispartof><rights>Author(s)</rights><rights>2021 Author(s). All article content, except where otherwise noted, is licensed under a Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).</rights><lds50>peer_reviewed</lds50><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c428t-c673b64a4db1c079ada351f4b6476590778f4e80b45fed39067003ae56fabfe43</citedby><cites>FETCH-LOGICAL-c428t-c673b64a4db1c079ada351f4b6476590778f4e80b45fed39067003ae56fabfe43</cites><orcidid>0000-0001-9052-3021</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>314,780,784,864,2100,27923,27924</link.rule.ids></links><search><creatorcontrib>Li, Chuanchang</creatorcontrib><creatorcontrib>Cui, Guohua</creatorcontrib><creatorcontrib>Zhang, Weiwei</creatorcontrib><creatorcontrib>Chen, Saixuan</creatorcontrib><creatorcontrib>Yang, Yinyi</creatorcontrib><title>Defect detection in vehicle mirror nonplanar surfaces with multi-scale atrous single-shot detect mechanism</title><title>AIP advances</title><description>Surface quality inspection is important for vehicle rearview mirrors. Surface quality defects, such as bubbles, particles, cracks, and scratches, may appear during the production process. Traditionally, manual inspection is time-consuming, laborious, low in accuracy, and costly. In recent years, object detectors based on deep learning have been used for defect detection [such as single shot detectors (SSDs)], and object detectors mostly detect the details of small objects at a shallow level. However, the lack of shallow semantic information will lead to inaccuracy. The deep layer has more semantic information, but the deep layer cannot be detected due to the effect of the complete convolutional layer in the convolutional neural network. In this article, we propose an enhanced SSD method to detect micro-defects on the nonplanar surface of vehicle rearview mirrors. We call it the multi-scale atrous single-shot detector (MSASSD). Specifically, we first replace the maximum pool depth layer with an unconscious convolutional layer to expand the receiving field without reducing the size of the input image. Then, we link the shallow layer to the deep layer through the fusion block to form new and rich fusion features for object detection. Finally, we use multi-scale features (including deep features and fusion features) to predict defects. The results show that our MSASSD method can improve the average accuracy of defect detection (about 1.2% compared with the SSD method), while the detection speed is equivalent (low about two frames per second compared to the SSD method).</description><subject>Artificial neural networks</subject><subject>Defects</subject><subject>Detectors</subject><subject>Flaw detection</subject><subject>Frames per second</subject><subject>Inspection</subject><subject>Object recognition</subject><subject>Semantics</subject><subject>Sensors</subject><subject>Surface properties</subject><issn>2158-3226</issn><issn>2158-3226</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2021</creationdate><recordtype>article</recordtype><sourceid>DOA</sourceid><recordid>eNp9kUtr3DAQgE1poSHJof9A0FMLTkZP28eSNG0g0Et7FiNZymqxra0kJ-TfR9tN00CguowYPr55Nc0HCmcUFD-XZwCS95K-aY4YlX3LGVNvX_zfN6c5b6E-MVDoxVGzvXTe2UJGV2oIcSFhIXduE-zkyBxSiokscdlNuGAieU0ercvkPpQNmdephDZbrCiWFNdMclhuJ9fmTfyrJLOzG1xCnk-adx6n7E6f4nHz6-rrz4vv7c2Pb9cXX25aK1hfWqs6bpRAMRpqoRtwRC6pFzXXKTlA1_VeuB6MkN6NfADVAXB0Unk03gl-3FwfvGPErd6lMGN60BGD_pOI6VZjKvsBNQIzZoTOeC6q0xg0auBM9d4CYwOvro8H1y7F36vLRW_jmpbavmZSdFQMA4dKfTpQNsWck_PPVSno_WW01E-XqeznA5ttKLjf-DN8F9M_UO9G_z_4tfkRmD-c1g</recordid><startdate>20210701</startdate><enddate>20210701</enddate><creator>Li, Chuanchang</creator><creator>Cui, Guohua</creator><creator>Zhang, Weiwei</creator><creator>Chen, Saixuan</creator><creator>Yang, Yinyi</creator><general>American Institute of Physics</general><general>AIP Publishing LLC</general><scope>AJDQP</scope><scope>AAYXX</scope><scope>CITATION</scope><scope>8FD</scope><scope>H8D</scope><scope>L7M</scope><scope>DOA</scope><orcidid>https://orcid.org/0000-0001-9052-3021</orcidid></search><sort><creationdate>20210701</creationdate><title>Defect detection in vehicle mirror nonplanar surfaces with multi-scale atrous single-shot detect mechanism</title><author>Li, Chuanchang ; Cui, Guohua ; Zhang, Weiwei ; Chen, Saixuan ; Yang, Yinyi</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c428t-c673b64a4db1c079ada351f4b6476590778f4e80b45fed39067003ae56fabfe43</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2021</creationdate><topic>Artificial neural networks</topic><topic>Defects</topic><topic>Detectors</topic><topic>Flaw detection</topic><topic>Frames per second</topic><topic>Inspection</topic><topic>Object recognition</topic><topic>Semantics</topic><topic>Sensors</topic><topic>Surface properties</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Li, Chuanchang</creatorcontrib><creatorcontrib>Cui, Guohua</creatorcontrib><creatorcontrib>Zhang, Weiwei</creatorcontrib><creatorcontrib>Chen, Saixuan</creatorcontrib><creatorcontrib>Yang, Yinyi</creatorcontrib><collection>AIP Open Access Journals</collection><collection>CrossRef</collection><collection>Technology Research Database</collection><collection>Aerospace Database</collection><collection>Advanced Technologies Database with Aerospace</collection><collection>DOAJ Directory of Open Access Journals</collection><jtitle>AIP advances</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Li, Chuanchang</au><au>Cui, Guohua</au><au>Zhang, Weiwei</au><au>Chen, Saixuan</au><au>Yang, Yinyi</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Defect detection in vehicle mirror nonplanar surfaces with multi-scale atrous single-shot detect mechanism</atitle><jtitle>AIP advances</jtitle><date>2021-07-01</date><risdate>2021</risdate><volume>11</volume><issue>7</issue><spage>075202</spage><epage>075202-11</epage><pages>075202-075202-11</pages><issn>2158-3226</issn><eissn>2158-3226</eissn><coden>AAIDBI</coden><abstract>Surface quality inspection is important for vehicle rearview mirrors. Surface quality defects, such as bubbles, particles, cracks, and scratches, may appear during the production process. Traditionally, manual inspection is time-consuming, laborious, low in accuracy, and costly. In recent years, object detectors based on deep learning have been used for defect detection [such as single shot detectors (SSDs)], and object detectors mostly detect the details of small objects at a shallow level. However, the lack of shallow semantic information will lead to inaccuracy. The deep layer has more semantic information, but the deep layer cannot be detected due to the effect of the complete convolutional layer in the convolutional neural network. In this article, we propose an enhanced SSD method to detect micro-defects on the nonplanar surface of vehicle rearview mirrors. We call it the multi-scale atrous single-shot detector (MSASSD). Specifically, we first replace the maximum pool depth layer with an unconscious convolutional layer to expand the receiving field without reducing the size of the input image. Then, we link the shallow layer to the deep layer through the fusion block to form new and rich fusion features for object detection. Finally, we use multi-scale features (including deep features and fusion features) to predict defects. The results show that our MSASSD method can improve the average accuracy of defect detection (about 1.2% compared with the SSD method), while the detection speed is equivalent (low about two frames per second compared to the SSD method).</abstract><cop>Melville</cop><pub>American Institute of Physics</pub><doi>10.1063/5.0053851</doi><tpages>11</tpages><orcidid>https://orcid.org/0000-0001-9052-3021</orcidid><oa>free_for_read</oa></addata></record> |
fulltext | fulltext |
identifier | ISSN: 2158-3226 |
ispartof | AIP advances, 2021-07, Vol.11 (7), p.075202-075202-11 |
issn | 2158-3226 2158-3226 |
language | eng |
recordid | cdi_proquest_journals_2547149930 |
source | DOAJ Directory of Open Access Journals; Elektronische Zeitschriftenbibliothek - Frei zugängliche E-Journals; Alma/SFX Local Collection; Free Full-Text Journals in Chemistry |
subjects | Artificial neural networks Defects Detectors Flaw detection Frames per second Inspection Object recognition Semantics Sensors Surface properties |
title | Defect detection in vehicle mirror nonplanar surfaces with multi-scale atrous single-shot detect mechanism |
url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-12T03%3A38%3A12IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_doaj_&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Defect%20detection%20in%20vehicle%20mirror%20nonplanar%20surfaces%20with%20multi-scale%20atrous%20single-shot%20detect%20mechanism&rft.jtitle=AIP%20advances&rft.au=Li,%20Chuanchang&rft.date=2021-07-01&rft.volume=11&rft.issue=7&rft.spage=075202&rft.epage=075202-11&rft.pages=075202-075202-11&rft.issn=2158-3226&rft.eissn=2158-3226&rft.coden=AAIDBI&rft_id=info:doi/10.1063/5.0053851&rft_dat=%3Cproquest_doaj_%3E2547149930%3C/proquest_doaj_%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2547149930&rft_id=info:pmid/&rft_doaj_id=oai_doaj_org_article_a02bbd07bf344e8bbab693268fc02293&rfr_iscdi=true |