Comparison of classic object-detection techniques for automated sewer defect detection
Sewer systems play a key role in cities to ensure public assets and safety. Timely detection of defects can effectively alleviate system deterioration. Conventional manual inspection is labor-intensive, error-prone and expensive. Object detection is a powerful deep learning technique that can comple...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IWA Publishing
2022-03-01
|
| Series: | Journal of Hydroinformatics |
| Subjects: | |
| Online Access: | http://jh.iwaponline.com/content/24/2/406 |
| _version_ | 1818009751310565376 |
|---|---|
| author | Qianqian Zhou Zuxiang Situ Shuai Teng Weifeng Chen Gongfa Chen Jiongheng Su |
| author_facet | Qianqian Zhou Zuxiang Situ Shuai Teng Weifeng Chen Gongfa Chen Jiongheng Su |
| author_sort | Qianqian Zhou |
| collection | DOAJ |
| description | Sewer systems play a key role in cities to ensure public assets and safety. Timely detection of defects can effectively alleviate system deterioration. Conventional manual inspection is labor-intensive, error-prone and expensive. Object detection is a powerful deep learning technique that can complement and/or replace conventional inspection, especially in complex environments. This study compares two classic object-detection methods, namely faster region-based convolutional neural network (R-CNN) and You Only Look Once (YOLO), for the detection and localization of five types of sewer defects. Model performances are evaluated based on their detection accuracy and processing speed under parameterization impacts of dataset size and training parameters. Results show that faster R-CNN achieved higher prediction accuracy. Training dataset size and maximum number of epochs (MaxE) had dominant impacts on model performances of faster R-CNN and YOLO, respectively. The processing speed increased along with the increasing training data for faster R-CNN, but did not vary significantly for YOLO. The models' abilities to detect disjoint and residential wall were highest, whereas crack and tree root were more difficult to detect. The results help to better understand the strengths and weaknesses of the classic methods and provide a useful user guidance for practical applications in automated sewer defect detection. HIGHLIGHTS
A deep learning technique for automated detection of multiple types of sewer defects.;
Compared the performances of two types of classic object-detection models.;
Evaluated model parameterization impacts and identification of key factors.; |
| first_indexed | 2024-04-14T05:46:52Z |
| format | Article |
| id | doaj.art-9cd423cce14f4832a527bc1632d2e676 |
| institution | Directory Open Access Journal |
| issn | 1464-7141 1465-1734 |
| language | English |
| last_indexed | 2024-04-14T05:46:52Z |
| publishDate | 2022-03-01 |
| publisher | IWA Publishing |
| record_format | Article |
| series | Journal of Hydroinformatics |
| spelling | doaj.art-9cd423cce14f4832a527bc1632d2e6762022-12-22T02:09:15ZengIWA PublishingJournal of Hydroinformatics1464-71411465-17342022-03-0124240641910.2166/hydro.2022.132132Comparison of classic object-detection techniques for automated sewer defect detectionQianqian Zhou0Zuxiang Situ1Shuai Teng2Weifeng Chen3Gongfa Chen4Jiongheng Su5 School of Civil and Transportation Engineering, Guangdong University of Technology, No. 100 Waihuan Xi Road, Guangzhou 510006, China School of Civil and Transportation Engineering, Guangdong University of Technology, No. 100 Waihuan Xi Road, Guangzhou 510006, China School of Civil and Transportation Engineering, Guangdong University of Technology, No. 100 Waihuan Xi Road, Guangzhou 510006, China School of Civil and Transportation Engineering, Guangdong University of Technology, No. 100 Waihuan Xi Road, Guangzhou 510006, China School of Civil and Transportation Engineering, Guangdong University of Technology, No. 100 Waihuan Xi Road, Guangzhou 510006, China Urban Development Research Center, Guangdong Urban & Rural Planning and Design Institute, No. 483 Nanzhou Road, Guangzhou 510290, China Sewer systems play a key role in cities to ensure public assets and safety. Timely detection of defects can effectively alleviate system deterioration. Conventional manual inspection is labor-intensive, error-prone and expensive. Object detection is a powerful deep learning technique that can complement and/or replace conventional inspection, especially in complex environments. This study compares two classic object-detection methods, namely faster region-based convolutional neural network (R-CNN) and You Only Look Once (YOLO), for the detection and localization of five types of sewer defects. Model performances are evaluated based on their detection accuracy and processing speed under parameterization impacts of dataset size and training parameters. Results show that faster R-CNN achieved higher prediction accuracy. Training dataset size and maximum number of epochs (MaxE) had dominant impacts on model performances of faster R-CNN and YOLO, respectively. The processing speed increased along with the increasing training data for faster R-CNN, but did not vary significantly for YOLO. The models' abilities to detect disjoint and residential wall were highest, whereas crack and tree root were more difficult to detect. The results help to better understand the strengths and weaknesses of the classic methods and provide a useful user guidance for practical applications in automated sewer defect detection. HIGHLIGHTS A deep learning technique for automated detection of multiple types of sewer defects.; Compared the performances of two types of classic object-detection models.; Evaluated model parameterization impacts and identification of key factors.;http://jh.iwaponline.com/content/24/2/406deep learningfaster r-cnnobject detectionsewer defect detectionyolo |
| spellingShingle | Qianqian Zhou Zuxiang Situ Shuai Teng Weifeng Chen Gongfa Chen Jiongheng Su Comparison of classic object-detection techniques for automated sewer defect detection Journal of Hydroinformatics deep learning faster r-cnn object detection sewer defect detection yolo |
| title | Comparison of classic object-detection techniques for automated sewer defect detection |
| title_full | Comparison of classic object-detection techniques for automated sewer defect detection |
| title_fullStr | Comparison of classic object-detection techniques for automated sewer defect detection |
| title_full_unstemmed | Comparison of classic object-detection techniques for automated sewer defect detection |
| title_short | Comparison of classic object-detection techniques for automated sewer defect detection |
| title_sort | comparison of classic object detection techniques for automated sewer defect detection |
| topic | deep learning faster r-cnn object detection sewer defect detection yolo |
| url | http://jh.iwaponline.com/content/24/2/406 |
| work_keys_str_mv | AT qianqianzhou comparisonofclassicobjectdetectiontechniquesforautomatedsewerdefectdetection AT zuxiangsitu comparisonofclassicobjectdetectiontechniquesforautomatedsewerdefectdetection AT shuaiteng comparisonofclassicobjectdetectiontechniquesforautomatedsewerdefectdetection AT weifengchen comparisonofclassicobjectdetectiontechniquesforautomatedsewerdefectdetection AT gongfachen comparisonofclassicobjectdetectiontechniquesforautomatedsewerdefectdetection AT jionghengsu comparisonofclassicobjectdetectiontechniquesforautomatedsewerdefectdetection |