Multi-Scale Polar Object Detection Based on Computer Vision
When ships navigate in polar regions, they may collide with ice masses, which may cause structural damage and endanger the safety of their occupants. Therefore, it is essential to promptly detect sea ice, icebergs, and passing ships. However, individual data sources have limits and should be combine...
Published in: | Water |
---|---|
Main Authors: | , , , , , |
Format: | Article in Journal/Newspaper |
Language: | English |
Published: |
MDPI AG
2023
|
Subjects: | |
Online Access: | https://doi.org/10.3390/w15193431 https://doaj.org/article/b3e45a195c7f4c14a70568aff353f6b6 |
id |
ftdoajarticles:oai:doaj.org/article:b3e45a195c7f4c14a70568aff353f6b6 |
---|---|
record_format |
openpolar |
spelling |
ftdoajarticles:oai:doaj.org/article:b3e45a195c7f4c14a70568aff353f6b6 2023-11-12T04:25:59+01:00 Multi-Scale Polar Object Detection Based on Computer Vision Shifeng Ding Dinghan Zeng Li Zhou Sen Han Fang Li Qingkai Wang 2023-09-01T00:00:00Z https://doi.org/10.3390/w15193431 https://doaj.org/article/b3e45a195c7f4c14a70568aff353f6b6 EN eng MDPI AG https://www.mdpi.com/2073-4441/15/19/3431 https://doaj.org/toc/2073-4441 doi:10.3390/w15193431 2073-4441 https://doaj.org/article/b3e45a195c7f4c14a70568aff353f6b6 Water, Vol 15, Iss 3431, p 3431 (2023) computer vision single-shot detector (SSD) You Only Look Once (YOLOv5) multi-source data polar object remote sensing image Hydraulic engineering TC1-978 Water supply for domestic and industrial purposes TD201-500 article 2023 ftdoajarticles https://doi.org/10.3390/w15193431 2023-10-15T00:35:10Z When ships navigate in polar regions, they may collide with ice masses, which may cause structural damage and endanger the safety of their occupants. Therefore, it is essential to promptly detect sea ice, icebergs, and passing ships. However, individual data sources have limits and should be combined and integrated to obtain more thorough information. A polar multi-target local-scale dataset with five categories was constructed. Sea ice, icebergs, ice melt ponds, icebreakers, and inter-ice channels were identified by a single-shot detector (SSD), with a final mAP value of 70.19%. A remote sensing sea ice dataset with 15,948 labels was constructed. The You Only Look Once (YOLOv5) model was improved with Squeeze-and-Excitation Networks (SE), Funnel Activation (FReLU), Fast Spatial Pyramid Pooling, and Cross Stage Partial Network (SPPCSPC-F). In the detection stage, a slicing operation was performed on remote sensing images to detect small targets. Simulated sea ice data were included to verify the model’s generalization ability. Then, the improved model was trained and evaluated in an ablation experiment. The mAP, recall (R), and precision (P) values of the improved YOLOv5 were 75.3%, 70.3, and 75.4%, with value increases of 3.5%, 3.4%, and 1.9%, respectively, compared to the original model. The improved YOLOv5 was also compared with other models such as YOLOv3, Faster-RCNN, and YOLOv4-tiny. The results indicated that the performance of the proposed model surpassed those of the other conventional models. This study achieved the detection of multiple targets on different scales in a polar region and realized data fusion, avoiding the limitations of using a single data source, and provides a method to support polar ship path planning. Article in Journal/Newspaper Sea ice Directory of Open Access Journals: DOAJ Articles Pyramid ENVELOPE(157.300,157.300,-81.333,-81.333) Water 15 19 3431 |
institution |
Open Polar |
collection |
Directory of Open Access Journals: DOAJ Articles |
op_collection_id |
ftdoajarticles |
language |
English |
topic |
computer vision single-shot detector (SSD) You Only Look Once (YOLOv5) multi-source data polar object remote sensing image Hydraulic engineering TC1-978 Water supply for domestic and industrial purposes TD201-500 |
spellingShingle |
computer vision single-shot detector (SSD) You Only Look Once (YOLOv5) multi-source data polar object remote sensing image Hydraulic engineering TC1-978 Water supply for domestic and industrial purposes TD201-500 Shifeng Ding Dinghan Zeng Li Zhou Sen Han Fang Li Qingkai Wang Multi-Scale Polar Object Detection Based on Computer Vision |
topic_facet |
computer vision single-shot detector (SSD) You Only Look Once (YOLOv5) multi-source data polar object remote sensing image Hydraulic engineering TC1-978 Water supply for domestic and industrial purposes TD201-500 |
description |
When ships navigate in polar regions, they may collide with ice masses, which may cause structural damage and endanger the safety of their occupants. Therefore, it is essential to promptly detect sea ice, icebergs, and passing ships. However, individual data sources have limits and should be combined and integrated to obtain more thorough information. A polar multi-target local-scale dataset with five categories was constructed. Sea ice, icebergs, ice melt ponds, icebreakers, and inter-ice channels were identified by a single-shot detector (SSD), with a final mAP value of 70.19%. A remote sensing sea ice dataset with 15,948 labels was constructed. The You Only Look Once (YOLOv5) model was improved with Squeeze-and-Excitation Networks (SE), Funnel Activation (FReLU), Fast Spatial Pyramid Pooling, and Cross Stage Partial Network (SPPCSPC-F). In the detection stage, a slicing operation was performed on remote sensing images to detect small targets. Simulated sea ice data were included to verify the model’s generalization ability. Then, the improved model was trained and evaluated in an ablation experiment. The mAP, recall (R), and precision (P) values of the improved YOLOv5 were 75.3%, 70.3, and 75.4%, with value increases of 3.5%, 3.4%, and 1.9%, respectively, compared to the original model. The improved YOLOv5 was also compared with other models such as YOLOv3, Faster-RCNN, and YOLOv4-tiny. The results indicated that the performance of the proposed model surpassed those of the other conventional models. This study achieved the detection of multiple targets on different scales in a polar region and realized data fusion, avoiding the limitations of using a single data source, and provides a method to support polar ship path planning. |
format |
Article in Journal/Newspaper |
author |
Shifeng Ding Dinghan Zeng Li Zhou Sen Han Fang Li Qingkai Wang |
author_facet |
Shifeng Ding Dinghan Zeng Li Zhou Sen Han Fang Li Qingkai Wang |
author_sort |
Shifeng Ding |
title |
Multi-Scale Polar Object Detection Based on Computer Vision |
title_short |
Multi-Scale Polar Object Detection Based on Computer Vision |
title_full |
Multi-Scale Polar Object Detection Based on Computer Vision |
title_fullStr |
Multi-Scale Polar Object Detection Based on Computer Vision |
title_full_unstemmed |
Multi-Scale Polar Object Detection Based on Computer Vision |
title_sort |
multi-scale polar object detection based on computer vision |
publisher |
MDPI AG |
publishDate |
2023 |
url |
https://doi.org/10.3390/w15193431 https://doaj.org/article/b3e45a195c7f4c14a70568aff353f6b6 |
long_lat |
ENVELOPE(157.300,157.300,-81.333,-81.333) |
geographic |
Pyramid |
geographic_facet |
Pyramid |
genre |
Sea ice |
genre_facet |
Sea ice |
op_source |
Water, Vol 15, Iss 3431, p 3431 (2023) |
op_relation |
https://www.mdpi.com/2073-4441/15/19/3431 https://doaj.org/toc/2073-4441 doi:10.3390/w15193431 2073-4441 https://doaj.org/article/b3e45a195c7f4c14a70568aff353f6b6 |
op_doi |
https://doi.org/10.3390/w15193431 |
container_title |
Water |
container_volume |
15 |
container_issue |
19 |
container_start_page |
3431 |
_version_ |
1782340127338528768 |