Loading…

Vehicle Detection in Aerial Images Based on Region Convolutional Neural Networks and Hard Negative Example Mining

Detecting vehicles in aerial imagery plays an important role in a wide range of applications. The current vehicle detection methods are mostly based on sliding-window search and handcrafted or shallow-learning-based features, having limited description capability and heavy computational costs. Recen...

Full description

Saved in:
Bibliographic Details
Published in:Sensors (Basel, Switzerland) Switzerland), 2017-02, Vol.17 (2), p.336-336
Main Authors: Tang, Tianyu, Zhou, Shilin, Deng, Zhipeng, Zou, Huanxin, Lei, Lin
Format: Article
Language:English
Subjects:
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
cited_by cdi_FETCH-LOGICAL-c601t-667827291550e75c0278a5263e33ffef7b9fd00973afa7f5991e994456077a283
cites cdi_FETCH-LOGICAL-c601t-667827291550e75c0278a5263e33ffef7b9fd00973afa7f5991e994456077a283
container_end_page 336
container_issue 2
container_start_page 336
container_title Sensors (Basel, Switzerland)
container_volume 17
creator Tang, Tianyu
Zhou, Shilin
Deng, Zhipeng
Zou, Huanxin
Lei, Lin
description Detecting vehicles in aerial imagery plays an important role in a wide range of applications. The current vehicle detection methods are mostly based on sliding-window search and handcrafted or shallow-learning-based features, having limited description capability and heavy computational costs. Recently, due to the powerful feature representations, region convolutional neural networks (CNN) based detection methods have achieved state-of-the-art performance in computer vision, especially Faster R-CNN. However, directly using it for vehicle detection in aerial images has many limitations: (1) region proposal network (RPN) in Faster R-CNN has poor performance for accurately locating small-sized vehicles, due to the relatively coarse feature maps; and (2) the classifier after RPN cannot distinguish vehicles and complex backgrounds well. In this study, an improved detection method based on Faster R-CNN is proposed in order to accomplish the two challenges mentioned above. Firstly, to improve the recall, we employ a hyper region proposal network (HRPN) to extract vehicle-like targets with a combination of hierarchical feature maps. Then, we replace the classifier after RPN by a cascade of boosted classifiers to verify the candidate regions, aiming at reducing false detection by negative example mining. We evaluate our method on the Munich vehicle dataset and the collected vehicle dataset, with improvements in accuracy and robustness compared to existing methods.
doi_str_mv 10.3390/s17020336
format article
fullrecord <record><control><sourceid>proquest_doaj_</sourceid><recordid>TN_cdi_doaj_primary_oai_doaj_org_article_aaa22e86438f4548b081712f7ed36092</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><doaj_id>oai_doaj_org_article_aaa22e86438f4548b081712f7ed36092</doaj_id><sourcerecordid>1869968385</sourcerecordid><originalsourceid>FETCH-LOGICAL-c601t-667827291550e75c0278a5263e33ffef7b9fd00973afa7f5991e994456077a283</originalsourceid><addsrcrecordid>eNqNkktv1DAUhSMEog9Y8AdQJDZ0MWD7xq8NUjsUOlIBCQFby5Ncpx4y8dROBvj3OJ0yatmAZOla93469rFPUTyj5BWAJq8TlYQRAPGgOKQVq2aKMfLwzv6gOEppRQgDAPW4OGCKEcWVPCyuv-GVrzss3-KA9eBDX_q-PMXobVcu1rbFVJ7ZhE2ZJ5-xnYB56LehGyc4Qx9xjDdl-BHi91TavikvbGxyp7WD32J5_tOuN_mID773ffukeORsl_DpbT0uvr47_zK_mF1-er-Yn17OakHoMBNCKiaZppwTlLwmTCrLmQAEcA6dXGrXEKIlWGel41pT1LqquCBSWqbguFjsdJtgV2YT_drGXyZYb24aIbbGxmHybqy1jKESFShX8UotiaKSMiexAUE0y1pvdlqbcbnGpsZ-yJ7vid6f9P7KtGFrOADXgmSBl7cCMVyPmAaz9qnGrrM9hjEZqjQoXUFe_0aVyq6JEP-BCq2FAsUz-uIvdBXGmP9voqSqoJJsuubJjqpjSCmi21ukxExRM_uoZfb53TfZk3-yBb8BycbLdQ</addsrcrecordid><sourcetype>Open Website</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>1878434720</pqid></control><display><type>article</type><title>Vehicle Detection in Aerial Images Based on Region Convolutional Neural Networks and Hard Negative Example Mining</title><source>Publicly Available Content Database</source><source>PubMed Central</source><creator>Tang, Tianyu ; Zhou, Shilin ; Deng, Zhipeng ; Zou, Huanxin ; Lei, Lin</creator><creatorcontrib>Tang, Tianyu ; Zhou, Shilin ; Deng, Zhipeng ; Zou, Huanxin ; Lei, Lin</creatorcontrib><description>Detecting vehicles in aerial imagery plays an important role in a wide range of applications. The current vehicle detection methods are mostly based on sliding-window search and handcrafted or shallow-learning-based features, having limited description capability and heavy computational costs. Recently, due to the powerful feature representations, region convolutional neural networks (CNN) based detection methods have achieved state-of-the-art performance in computer vision, especially Faster R-CNN. However, directly using it for vehicle detection in aerial images has many limitations: (1) region proposal network (RPN) in Faster R-CNN has poor performance for accurately locating small-sized vehicles, due to the relatively coarse feature maps; and (2) the classifier after RPN cannot distinguish vehicles and complex backgrounds well. In this study, an improved detection method based on Faster R-CNN is proposed in order to accomplish the two challenges mentioned above. Firstly, to improve the recall, we employ a hyper region proposal network (HRPN) to extract vehicle-like targets with a combination of hierarchical feature maps. Then, we replace the classifier after RPN by a cascade of boosted classifiers to verify the candidate regions, aiming at reducing false detection by negative example mining. We evaluate our method on the Munich vehicle dataset and the collected vehicle dataset, with improvements in accuracy and robustness compared to existing methods.</description><identifier>ISSN: 1424-8220</identifier><identifier>EISSN: 1424-8220</identifier><identifier>DOI: 10.3390/s17020336</identifier><identifier>PMID: 28208587</identifier><language>eng</language><publisher>Switzerland: MDPI AG</publisher><subject>Aerials ; Classifiers ; Computer vision ; convolutional neural networks ; Feature maps ; hard negative example mining ; hyper region proposal network ; Image detection ; Networks ; Neural networks ; Proposals ; vehicle detection ; Vehicles</subject><ispartof>Sensors (Basel, Switzerland), 2017-02, Vol.17 (2), p.336-336</ispartof><rights>2017. This work is licensed under https://creativecommons.org/licenses/by/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.</rights><rights>2017 by the authors. 2017</rights><lds50>peer_reviewed</lds50><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c601t-667827291550e75c0278a5263e33ffef7b9fd00973afa7f5991e994456077a283</citedby><cites>FETCH-LOGICAL-c601t-667827291550e75c0278a5263e33ffef7b9fd00973afa7f5991e994456077a283</cites><orcidid>0000-0002-9986-1681</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktopdf>$$Uhttps://www.proquest.com/docview/1878434720/fulltextPDF?pq-origsite=primo$$EPDF$$P50$$Gproquest$$Hfree_for_read</linktopdf><linktohtml>$$Uhttps://www.proquest.com/docview/1878434720?pq-origsite=primo$$EHTML$$P50$$Gproquest$$Hfree_for_read</linktohtml><link.rule.ids>230,314,727,780,784,885,25753,27924,27925,37012,37013,44590,53791,53793,75126</link.rule.ids><backlink>$$Uhttps://www.ncbi.nlm.nih.gov/pubmed/28208587$$D View this record in MEDLINE/PubMed$$Hfree_for_read</backlink></links><search><creatorcontrib>Tang, Tianyu</creatorcontrib><creatorcontrib>Zhou, Shilin</creatorcontrib><creatorcontrib>Deng, Zhipeng</creatorcontrib><creatorcontrib>Zou, Huanxin</creatorcontrib><creatorcontrib>Lei, Lin</creatorcontrib><title>Vehicle Detection in Aerial Images Based on Region Convolutional Neural Networks and Hard Negative Example Mining</title><title>Sensors (Basel, Switzerland)</title><addtitle>Sensors (Basel)</addtitle><description>Detecting vehicles in aerial imagery plays an important role in a wide range of applications. The current vehicle detection methods are mostly based on sliding-window search and handcrafted or shallow-learning-based features, having limited description capability and heavy computational costs. Recently, due to the powerful feature representations, region convolutional neural networks (CNN) based detection methods have achieved state-of-the-art performance in computer vision, especially Faster R-CNN. However, directly using it for vehicle detection in aerial images has many limitations: (1) region proposal network (RPN) in Faster R-CNN has poor performance for accurately locating small-sized vehicles, due to the relatively coarse feature maps; and (2) the classifier after RPN cannot distinguish vehicles and complex backgrounds well. In this study, an improved detection method based on Faster R-CNN is proposed in order to accomplish the two challenges mentioned above. Firstly, to improve the recall, we employ a hyper region proposal network (HRPN) to extract vehicle-like targets with a combination of hierarchical feature maps. Then, we replace the classifier after RPN by a cascade of boosted classifiers to verify the candidate regions, aiming at reducing false detection by negative example mining. We evaluate our method on the Munich vehicle dataset and the collected vehicle dataset, with improvements in accuracy and robustness compared to existing methods.</description><subject>Aerials</subject><subject>Classifiers</subject><subject>Computer vision</subject><subject>convolutional neural networks</subject><subject>Feature maps</subject><subject>hard negative example mining</subject><subject>hyper region proposal network</subject><subject>Image detection</subject><subject>Networks</subject><subject>Neural networks</subject><subject>Proposals</subject><subject>vehicle detection</subject><subject>Vehicles</subject><issn>1424-8220</issn><issn>1424-8220</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2017</creationdate><recordtype>article</recordtype><sourceid>PIMPY</sourceid><sourceid>DOA</sourceid><recordid>eNqNkktv1DAUhSMEog9Y8AdQJDZ0MWD7xq8NUjsUOlIBCQFby5Ncpx4y8dROBvj3OJ0yatmAZOla93469rFPUTyj5BWAJq8TlYQRAPGgOKQVq2aKMfLwzv6gOEppRQgDAPW4OGCKEcWVPCyuv-GVrzss3-KA9eBDX_q-PMXobVcu1rbFVJ7ZhE2ZJ5-xnYB56LehGyc4Qx9xjDdl-BHi91TavikvbGxyp7WD32J5_tOuN_mID773ffukeORsl_DpbT0uvr47_zK_mF1-er-Yn17OakHoMBNCKiaZppwTlLwmTCrLmQAEcA6dXGrXEKIlWGel41pT1LqquCBSWqbguFjsdJtgV2YT_drGXyZYb24aIbbGxmHybqy1jKESFShX8UotiaKSMiexAUE0y1pvdlqbcbnGpsZ-yJ7vid6f9P7KtGFrOADXgmSBl7cCMVyPmAaz9qnGrrM9hjEZqjQoXUFe_0aVyq6JEP-BCq2FAsUz-uIvdBXGmP9voqSqoJJsuubJjqpjSCmi21ukxExRM_uoZfb53TfZk3-yBb8BycbLdQ</recordid><startdate>20170210</startdate><enddate>20170210</enddate><creator>Tang, Tianyu</creator><creator>Zhou, Shilin</creator><creator>Deng, Zhipeng</creator><creator>Zou, Huanxin</creator><creator>Lei, Lin</creator><general>MDPI AG</general><general>MDPI</general><scope>NPM</scope><scope>AAYXX</scope><scope>CITATION</scope><scope>3V.</scope><scope>7X7</scope><scope>7XB</scope><scope>88E</scope><scope>8FI</scope><scope>8FJ</scope><scope>8FK</scope><scope>ABUWG</scope><scope>AFKRA</scope><scope>AZQEC</scope><scope>BENPR</scope><scope>CCPQU</scope><scope>DWQXO</scope><scope>FYUFA</scope><scope>GHDGH</scope><scope>K9.</scope><scope>M0S</scope><scope>M1P</scope><scope>PIMPY</scope><scope>PQEST</scope><scope>PQQKQ</scope><scope>PQUKI</scope><scope>PRINS</scope><scope>7X8</scope><scope>7QO</scope><scope>8FD</scope><scope>FR3</scope><scope>P64</scope><scope>7SP</scope><scope>7TB</scope><scope>7U5</scope><scope>L7M</scope><scope>5PM</scope><scope>DOA</scope><orcidid>https://orcid.org/0000-0002-9986-1681</orcidid></search><sort><creationdate>20170210</creationdate><title>Vehicle Detection in Aerial Images Based on Region Convolutional Neural Networks and Hard Negative Example Mining</title><author>Tang, Tianyu ; Zhou, Shilin ; Deng, Zhipeng ; Zou, Huanxin ; Lei, Lin</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c601t-667827291550e75c0278a5263e33ffef7b9fd00973afa7f5991e994456077a283</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2017</creationdate><topic>Aerials</topic><topic>Classifiers</topic><topic>Computer vision</topic><topic>convolutional neural networks</topic><topic>Feature maps</topic><topic>hard negative example mining</topic><topic>hyper region proposal network</topic><topic>Image detection</topic><topic>Networks</topic><topic>Neural networks</topic><topic>Proposals</topic><topic>vehicle detection</topic><topic>Vehicles</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Tang, Tianyu</creatorcontrib><creatorcontrib>Zhou, Shilin</creatorcontrib><creatorcontrib>Deng, Zhipeng</creatorcontrib><creatorcontrib>Zou, Huanxin</creatorcontrib><creatorcontrib>Lei, Lin</creatorcontrib><collection>PubMed</collection><collection>CrossRef</collection><collection>ProQuest Central (Corporate)</collection><collection>ProQuest Health and Medical</collection><collection>ProQuest Central (purchase pre-March 2016)</collection><collection>Medical Database (Alumni Edition)</collection><collection>Hospital Premium Collection</collection><collection>Hospital Premium Collection (Alumni Edition)</collection><collection>ProQuest Central (Alumni) (purchase pre-March 2016)</collection><collection>ProQuest Central (Alumni)</collection><collection>ProQuest Central</collection><collection>ProQuest Central Essentials</collection><collection>AUTh Library subscriptions: ProQuest Central</collection><collection>ProQuest One Community College</collection><collection>ProQuest Central Korea</collection><collection>Health Research Premium Collection</collection><collection>Health Research Premium Collection (Alumni)</collection><collection>ProQuest Health &amp; Medical Complete (Alumni)</collection><collection>Health &amp; Medical Collection (Alumni Edition)</collection><collection>PML(ProQuest Medical Library)</collection><collection>Publicly Available Content Database</collection><collection>ProQuest One Academic Eastern Edition (DO NOT USE)</collection><collection>ProQuest One Academic</collection><collection>ProQuest One Academic UKI Edition</collection><collection>ProQuest Central China</collection><collection>MEDLINE - Academic</collection><collection>Biotechnology Research Abstracts</collection><collection>Technology Research Database</collection><collection>Engineering Research Database</collection><collection>Biotechnology and BioEngineering Abstracts</collection><collection>Electronics &amp; Communications Abstracts</collection><collection>Mechanical &amp; Transportation Engineering Abstracts</collection><collection>Solid State and Superconductivity Abstracts</collection><collection>Advanced Technologies Database with Aerospace</collection><collection>PubMed Central (Full Participant titles)</collection><collection>Directory of Open Access Journals</collection><jtitle>Sensors (Basel, Switzerland)</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Tang, Tianyu</au><au>Zhou, Shilin</au><au>Deng, Zhipeng</au><au>Zou, Huanxin</au><au>Lei, Lin</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Vehicle Detection in Aerial Images Based on Region Convolutional Neural Networks and Hard Negative Example Mining</atitle><jtitle>Sensors (Basel, Switzerland)</jtitle><addtitle>Sensors (Basel)</addtitle><date>2017-02-10</date><risdate>2017</risdate><volume>17</volume><issue>2</issue><spage>336</spage><epage>336</epage><pages>336-336</pages><issn>1424-8220</issn><eissn>1424-8220</eissn><abstract>Detecting vehicles in aerial imagery plays an important role in a wide range of applications. The current vehicle detection methods are mostly based on sliding-window search and handcrafted or shallow-learning-based features, having limited description capability and heavy computational costs. Recently, due to the powerful feature representations, region convolutional neural networks (CNN) based detection methods have achieved state-of-the-art performance in computer vision, especially Faster R-CNN. However, directly using it for vehicle detection in aerial images has many limitations: (1) region proposal network (RPN) in Faster R-CNN has poor performance for accurately locating small-sized vehicles, due to the relatively coarse feature maps; and (2) the classifier after RPN cannot distinguish vehicles and complex backgrounds well. In this study, an improved detection method based on Faster R-CNN is proposed in order to accomplish the two challenges mentioned above. Firstly, to improve the recall, we employ a hyper region proposal network (HRPN) to extract vehicle-like targets with a combination of hierarchical feature maps. Then, we replace the classifier after RPN by a cascade of boosted classifiers to verify the candidate regions, aiming at reducing false detection by negative example mining. We evaluate our method on the Munich vehicle dataset and the collected vehicle dataset, with improvements in accuracy and robustness compared to existing methods.</abstract><cop>Switzerland</cop><pub>MDPI AG</pub><pmid>28208587</pmid><doi>10.3390/s17020336</doi><tpages>1</tpages><orcidid>https://orcid.org/0000-0002-9986-1681</orcidid><oa>free_for_read</oa></addata></record>
fulltext fulltext
identifier ISSN: 1424-8220
ispartof Sensors (Basel, Switzerland), 2017-02, Vol.17 (2), p.336-336
issn 1424-8220
1424-8220
language eng
recordid cdi_doaj_primary_oai_doaj_org_article_aaa22e86438f4548b081712f7ed36092
source Publicly Available Content Database; PubMed Central
subjects Aerials
Classifiers
Computer vision
convolutional neural networks
Feature maps
hard negative example mining
hyper region proposal network
Image detection
Networks
Neural networks
Proposals
vehicle detection
Vehicles
title Vehicle Detection in Aerial Images Based on Region Convolutional Neural Networks and Hard Negative Example Mining
url http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-24T10%3A13%3A46IST&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=Vehicle%20Detection%20in%20Aerial%20Images%20Based%20on%20Region%20Convolutional%20Neural%20Networks%20and%20Hard%20Negative%20Example%20Mining&rft.jtitle=Sensors%20(Basel,%20Switzerland)&rft.au=Tang,%20Tianyu&rft.date=2017-02-10&rft.volume=17&rft.issue=2&rft.spage=336&rft.epage=336&rft.pages=336-336&rft.issn=1424-8220&rft.eissn=1424-8220&rft_id=info:doi/10.3390/s17020336&rft_dat=%3Cproquest_doaj_%3E1869968385%3C/proquest_doaj_%3E%3Cgrp_id%3Ecdi_FETCH-LOGICAL-c601t-667827291550e75c0278a5263e33ffef7b9fd00973afa7f5991e994456077a283%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_pqid=1878434720&rft_id=info:pmid/28208587&rfr_iscdi=true