Detecting Vietnam War bomb craters in declassified historical KH-9 satellite imagery
Thousands of people are injured every year from explosive remnants of war which include unexploded ordnance (UXO) and abandoned ordnance. UXO has negative long-term impacts on livelihoods and ecosystems in contaminated areas. Exact locations of remaining UXO are often unknown as survey and clearance...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2024-12-01
|
| Series: | Science of Remote Sensing |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2666017224000270 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850060769245265920 |
|---|---|
| author | Philipp Barthelme Eoghan Darbyshire Dominick V. Spracklen Gary R. Watmough |
| author_facet | Philipp Barthelme Eoghan Darbyshire Dominick V. Spracklen Gary R. Watmough |
| author_sort | Philipp Barthelme |
| collection | DOAJ |
| description | Thousands of people are injured every year from explosive remnants of war which include unexploded ordnance (UXO) and abandoned ordnance. UXO has negative long-term impacts on livelihoods and ecosystems in contaminated areas. Exact locations of remaining UXO are often unknown as survey and clearance activities can be dangerous, expensive and time-consuming. In Vietnam, Lao PDR and Cambodia, about 20% of the land remains contaminated by UXO from the Vietnam War. Recently declassified historical KH-9 satellite imagery, taken during and immediately after the Vietnam War, now provides an opportunity to map this remaining contamination. KH-9 imagery was acquired and orthorectified for two study areas in Southeast Asia. Bomb craters were manually labeled in a subset of the imagery to train convolutional neural networks (CNNs) for automated crater detection. The CNNs achieved a F1-Score of 0.61 and identified more than 500,000 bomb craters across the two study areas. The detected craters provided more precise information on the impact locations of bombs than target locations available from declassified U.S. bombing records. This could allow for a more precise localization of suspected hazardous areas during non-technical surveys as well as a more fine-grained determination of residual risk of UXO. The method is directly transferable to other areas in Southeast Asia and is cost-effective due to the low cost of the KH-9 imagery and the use of open-source software. The results also show the potential of integrating crater detection into data-driven decision making in mine action across more recent conflicts. |
| format | Article |
| id | doaj-art-c2706bc3e4dd4b7aa4177b2d14e64574 |
| institution | DOAJ |
| issn | 2666-0172 |
| language | English |
| publishDate | 2024-12-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Science of Remote Sensing |
| spelling | doaj-art-c2706bc3e4dd4b7aa4177b2d14e645742025-08-20T02:50:27ZengElsevierScience of Remote Sensing2666-01722024-12-011010014310.1016/j.srs.2024.100143Detecting Vietnam War bomb craters in declassified historical KH-9 satellite imageryPhilipp Barthelme0Eoghan Darbyshire1Dominick V. Spracklen2Gary R. Watmough3School of Geosciences, University of Edinburgh, Edinburgh, United Kingdom; Corresponding author.Conflict and Environment Observatory, Mytholmroyd, United KingdomSchool of Earth and Environment, University of Leeds, Leeds, United KingdomSchool of Geosciences, University of Edinburgh, Edinburgh, United Kingdom; Global Academy of Agriculture and Food Systems, University of Edinburgh, United Kingdom; Novel Data Ecosystems for Sustainability (NODES) Research Group. Advancing Systems Analysis Program, International Institute for Applied Systems Analysis (IIASA), AustriaThousands of people are injured every year from explosive remnants of war which include unexploded ordnance (UXO) and abandoned ordnance. UXO has negative long-term impacts on livelihoods and ecosystems in contaminated areas. Exact locations of remaining UXO are often unknown as survey and clearance activities can be dangerous, expensive and time-consuming. In Vietnam, Lao PDR and Cambodia, about 20% of the land remains contaminated by UXO from the Vietnam War. Recently declassified historical KH-9 satellite imagery, taken during and immediately after the Vietnam War, now provides an opportunity to map this remaining contamination. KH-9 imagery was acquired and orthorectified for two study areas in Southeast Asia. Bomb craters were manually labeled in a subset of the imagery to train convolutional neural networks (CNNs) for automated crater detection. The CNNs achieved a F1-Score of 0.61 and identified more than 500,000 bomb craters across the two study areas. The detected craters provided more precise information on the impact locations of bombs than target locations available from declassified U.S. bombing records. This could allow for a more precise localization of suspected hazardous areas during non-technical surveys as well as a more fine-grained determination of residual risk of UXO. The method is directly transferable to other areas in Southeast Asia and is cost-effective due to the low cost of the KH-9 imagery and the use of open-source software. The results also show the potential of integrating crater detection into data-driven decision making in mine action across more recent conflicts.http://www.sciencedirect.com/science/article/pii/S2666017224000270Unexploded ordnanceBomb craterSatellite imageryKH-9 HexagonVietnam WarMine action |
| spellingShingle | Philipp Barthelme Eoghan Darbyshire Dominick V. Spracklen Gary R. Watmough Detecting Vietnam War bomb craters in declassified historical KH-9 satellite imagery Science of Remote Sensing Unexploded ordnance Bomb crater Satellite imagery KH-9 Hexagon Vietnam War Mine action |
| title | Detecting Vietnam War bomb craters in declassified historical KH-9 satellite imagery |
| title_full | Detecting Vietnam War bomb craters in declassified historical KH-9 satellite imagery |
| title_fullStr | Detecting Vietnam War bomb craters in declassified historical KH-9 satellite imagery |
| title_full_unstemmed | Detecting Vietnam War bomb craters in declassified historical KH-9 satellite imagery |
| title_short | Detecting Vietnam War bomb craters in declassified historical KH-9 satellite imagery |
| title_sort | detecting vietnam war bomb craters in declassified historical kh 9 satellite imagery |
| topic | Unexploded ordnance Bomb crater Satellite imagery KH-9 Hexagon Vietnam War Mine action |
| url | http://www.sciencedirect.com/science/article/pii/S2666017224000270 |
| work_keys_str_mv | AT philippbarthelme detectingvietnamwarbombcratersindeclassifiedhistoricalkh9satelliteimagery AT eoghandarbyshire detectingvietnamwarbombcratersindeclassifiedhistoricalkh9satelliteimagery AT dominickvspracklen detectingvietnamwarbombcratersindeclassifiedhistoricalkh9satelliteimagery AT garyrwatmough detectingvietnamwarbombcratersindeclassifiedhistoricalkh9satelliteimagery |