Storm-Drain and Manhole Detection Using the RetinaNet Method
As key-components of the urban-drainage system, storm-drains and manholes are essential to the hydrological modeling of urban basins. Accurately mapping of these objects can help to improve the storm-drain systems for the prevention and mitigation of urban floods. Novel Deep Learning (DL) methods ha...
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-08-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/20/16/4450 |
| _version_ | 1797559228504735744 |
|---|---|
| author | Anderson Santos José Marcato Junior Jonathan de Andrade Silva Rodrigo Pereira Daniel Matos Geazy Menezes Leandro Higa Anette Eltner Ana Paula Ramos Lucas Osco Wesley Gonçalves |
| author_facet | Anderson Santos José Marcato Junior Jonathan de Andrade Silva Rodrigo Pereira Daniel Matos Geazy Menezes Leandro Higa Anette Eltner Ana Paula Ramos Lucas Osco Wesley Gonçalves |
| author_sort | Anderson Santos |
| collection | DOAJ |
| description | As key-components of the urban-drainage system, storm-drains and manholes are essential to the hydrological modeling of urban basins. Accurately mapping of these objects can help to improve the storm-drain systems for the prevention and mitigation of urban floods. Novel Deep Learning (DL) methods have been proposed to aid the mapping of these urban features. The main aim of this paper is to evaluate the state-of-the-art object detection method RetinaNet to identify storm-drain and manhole in urban areas in street-level RGB images. The experimental assessment was performed using 297 mobile mapping images captured in 2019 in the streets in six regions in Campo Grande city, located in Mato Grosso do Sul state, Brazil. Two configurations of training, validation, and test images were considered. ResNet-50 and ResNet-101 were adopted in the experimental assessment as the two distinct feature extractor networks (i.e., backbones) for the RetinaNet method. The results were compared with the Faster R-CNN method. The results showed a higher detection accuracy when using RetinaNet with ResNet-50. In conclusion, the assessed DL method is adequate to detect storm-drain and manhole from mobile mapping RGB images, outperforming the Faster R-CNN method. The labeled dataset used in this study is available for future research. |
| first_indexed | 2024-03-10T17:42:27Z |
| format | Article |
| id | doaj.art-0fa9c85779a640068d6a29b46296aaab |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-10T17:42:27Z |
| publishDate | 2020-08-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-0fa9c85779a640068d6a29b46296aaab2023-11-20T09:37:58ZengMDPI AGSensors1424-82202020-08-012016445010.3390/s20164450Storm-Drain and Manhole Detection Using the RetinaNet MethodAnderson Santos0José Marcato Junior1Jonathan de Andrade Silva2Rodrigo Pereira3Daniel Matos4Geazy Menezes5Leandro Higa6Anette Eltner7Ana Paula Ramos8Lucas Osco9Wesley Gonçalves10Faculty of Computer Science, Federal University of Mato Grosso do Sul, Campo Grande 79070900, MS, BrazilFaculty of Engineering, Architecture and Urbanism and Geography, Federal University of Mato Grosso do Sul, Campo Grande 79070900, MS, BrazilFaculty of Computer Science, Federal University of Mato Grosso do Sul, Campo Grande 79070900, MS, BrazilFaculty of Engineering, Architecture and Urbanism and Geography, Federal University of Mato Grosso do Sul, Campo Grande 79070900, MS, BrazilFaculty of Engineering, Architecture and Urbanism and Geography, Federal University of Mato Grosso do Sul, Campo Grande 79070900, MS, BrazilFaculty of Computer Science, Federal University of Mato Grosso do Sul, Campo Grande 79070900, MS, BrazilFaculty of Computer Science, Federal University of Mato Grosso do Sul, Campo Grande 79070900, MS, BrazilInstitute of Photogrammetry and Remote Sensing, Technische Universität Dresden, 01062 Dresden, GermanyGraduate Program of Environment and Regional Development, University of Western São Paulo, Presidente Prudente 19067175, BrazilGraduate Program of Environment and Regional Development, University of Western São Paulo, Presidente Prudente 19067175, BrazilFaculty of Computer Science, Federal University of Mato Grosso do Sul, Campo Grande 79070900, MS, BrazilAs key-components of the urban-drainage system, storm-drains and manholes are essential to the hydrological modeling of urban basins. Accurately mapping of these objects can help to improve the storm-drain systems for the prevention and mitigation of urban floods. Novel Deep Learning (DL) methods have been proposed to aid the mapping of these urban features. The main aim of this paper is to evaluate the state-of-the-art object detection method RetinaNet to identify storm-drain and manhole in urban areas in street-level RGB images. The experimental assessment was performed using 297 mobile mapping images captured in 2019 in the streets in six regions in Campo Grande city, located in Mato Grosso do Sul state, Brazil. Two configurations of training, validation, and test images were considered. ResNet-50 and ResNet-101 were adopted in the experimental assessment as the two distinct feature extractor networks (i.e., backbones) for the RetinaNet method. The results were compared with the Faster R-CNN method. The results showed a higher detection accuracy when using RetinaNet with ResNet-50. In conclusion, the assessed DL method is adequate to detect storm-drain and manhole from mobile mapping RGB images, outperforming the Faster R-CNN method. The labeled dataset used in this study is available for future research.https://www.mdpi.com/1424-8220/20/16/4450convolutional neural networkobject detectionurban floods mapping |
| spellingShingle | Anderson Santos José Marcato Junior Jonathan de Andrade Silva Rodrigo Pereira Daniel Matos Geazy Menezes Leandro Higa Anette Eltner Ana Paula Ramos Lucas Osco Wesley Gonçalves Storm-Drain and Manhole Detection Using the RetinaNet Method Sensors convolutional neural network object detection urban floods mapping |
| title | Storm-Drain and Manhole Detection Using the RetinaNet Method |
| title_full | Storm-Drain and Manhole Detection Using the RetinaNet Method |
| title_fullStr | Storm-Drain and Manhole Detection Using the RetinaNet Method |
| title_full_unstemmed | Storm-Drain and Manhole Detection Using the RetinaNet Method |
| title_short | Storm-Drain and Manhole Detection Using the RetinaNet Method |
| title_sort | storm drain and manhole detection using the retinanet method |
| topic | convolutional neural network object detection urban floods mapping |
| url | https://www.mdpi.com/1424-8220/20/16/4450 |
| work_keys_str_mv | AT andersonsantos stormdrainandmanholedetectionusingtheretinanetmethod AT josemarcatojunior stormdrainandmanholedetectionusingtheretinanetmethod AT jonathandeandradesilva stormdrainandmanholedetectionusingtheretinanetmethod AT rodrigopereira stormdrainandmanholedetectionusingtheretinanetmethod AT danielmatos stormdrainandmanholedetectionusingtheretinanetmethod AT geazymenezes stormdrainandmanholedetectionusingtheretinanetmethod AT leandrohiga stormdrainandmanholedetectionusingtheretinanetmethod AT anetteeltner stormdrainandmanholedetectionusingtheretinanetmethod AT anapaularamos stormdrainandmanholedetectionusingtheretinanetmethod AT lucasosco stormdrainandmanholedetectionusingtheretinanetmethod AT wesleygoncalves stormdrainandmanholedetectionusingtheretinanetmethod |