Few-Shot Contrail Segmentation in Remote Sensing Imagery With Loss Function in Hough Space
Condensation trails, or contrails, are line-shaped clouds that are produced by an aircraft engine exhaust. These contrails often impact climate significantly due to their potential warming effect. Identification of contrail formation through satellite images has been an ongoing research challenge. T...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2025-01-01
|
| Series: | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10820969/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832576761101025280 |
|---|---|
| author | Junzi Sun Esther Roosenbrand |
| author_facet | Junzi Sun Esther Roosenbrand |
| author_sort | Junzi Sun |
| collection | DOAJ |
| description | Condensation trails, or contrails, are line-shaped clouds that are produced by an aircraft engine exhaust. These contrails often impact climate significantly due to their potential warming effect. Identification of contrail formation through satellite images has been an ongoing research challenge. Traditional computer vision techniques struggle with varying imagery conditions, and supervised machine learning approaches often require a large amount of hand-labeled images. This study researches few-shot transfer learning and provides an innovative approach for contrail segmentation with a few labeled images. The methodology leverages backbone segmentation models, which are pretrained on existing image datasets and fine-tuned using an augmented contrail-specific dataset. We also introduce a new loss function, SR loss, which enhances contrail line detection by incorporating Hough transformation in model training. This transformation improves performance over generic image segmentation loss functions. The openly shared few-shot learning library, contrail-seg, has demonstrated that few-shot learning can be effectively applied to contrail segmentation with the new loss function. |
| format | Article |
| id | doaj-art-69bdbae0cdeb4269b6083494cb89c452 |
| institution | Kabale University |
| issn | 1939-1404 2151-1535 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| spelling | doaj-art-69bdbae0cdeb4269b6083494cb89c4522025-01-31T00:00:20ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-01184273428510.1109/JSTARS.2025.352557610820969Few-Shot Contrail Segmentation in Remote Sensing Imagery With Loss Function in Hough SpaceJunzi Sun0https://orcid.org/0000-0003-3888-1192Esther Roosenbrand1Faculty of Aerospace Engineering, Delft University of Technology Delft, Delft, The NetherlandsFaculty of Aerospace Engineering, Delft University of Technology Delft, Delft, The NetherlandsCondensation trails, or contrails, are line-shaped clouds that are produced by an aircraft engine exhaust. These contrails often impact climate significantly due to their potential warming effect. Identification of contrail formation through satellite images has been an ongoing research challenge. Traditional computer vision techniques struggle with varying imagery conditions, and supervised machine learning approaches often require a large amount of hand-labeled images. This study researches few-shot transfer learning and provides an innovative approach for contrail segmentation with a few labeled images. The methodology leverages backbone segmentation models, which are pretrained on existing image datasets and fine-tuned using an augmented contrail-specific dataset. We also introduce a new loss function, SR loss, which enhances contrail line detection by incorporating Hough transformation in model training. This transformation improves performance over generic image segmentation loss functions. The openly shared few-shot learning library, contrail-seg, has demonstrated that few-shot learning can be effectively applied to contrail segmentation with the new loss function.https://ieeexplore.ieee.org/document/10820969/Contrail detectioncontrail segmentationfew-shot learningremote sensingSR loss |
| spellingShingle | Junzi Sun Esther Roosenbrand Few-Shot Contrail Segmentation in Remote Sensing Imagery With Loss Function in Hough Space IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing Contrail detection contrail segmentation few-shot learning remote sensing SR loss |
| title | Few-Shot Contrail Segmentation in Remote Sensing Imagery With Loss Function in Hough Space |
| title_full | Few-Shot Contrail Segmentation in Remote Sensing Imagery With Loss Function in Hough Space |
| title_fullStr | Few-Shot Contrail Segmentation in Remote Sensing Imagery With Loss Function in Hough Space |
| title_full_unstemmed | Few-Shot Contrail Segmentation in Remote Sensing Imagery With Loss Function in Hough Space |
| title_short | Few-Shot Contrail Segmentation in Remote Sensing Imagery With Loss Function in Hough Space |
| title_sort | few shot contrail segmentation in remote sensing imagery with loss function in hough space |
| topic | Contrail detection contrail segmentation few-shot learning remote sensing SR loss |
| url | https://ieeexplore.ieee.org/document/10820969/ |
| work_keys_str_mv | AT junzisun fewshotcontrailsegmentationinremotesensingimagerywithlossfunctioninhoughspace AT estherroosenbrand fewshotcontrailsegmentationinremotesensingimagerywithlossfunctioninhoughspace |