VedgeSat looks at publicly available satellite imagery to identify the edge of the plant, the line closest to the water surface where the plant begins to grow. uses machine learning Classify each pixel as either “vegetated” or “non-vegetated,” allowing vegetation lines to be established across different sites. This tool can provide insight into how coastal erosion is progressing by monitoring how that line changes over time.
The software was first used and monitored on Scottish coasts in 2024. Vegetation edge of St Andrews and Eden Estuary. The Glasgow team has now successfully tested it in a tropical environment, demonstrating its capabilities in a completely different coastal landscape. The works are published in Frontiers of marine science.
“This study provides the first benchmark for automatic vegetation edge detection in tropical systems and demonstrates the scalability of satellite-based coastal monitoring,” said corresponding author Idam Nugraha from the Glasgow School of Geography and Earth Sciences and the Islamic Riau University in Indonesia.
“VedgeSat has the potential to support coastal management in data-poor areas around the world by enabling large-scale, reproducible, and cost-effective analysis.”
The researchers used VedgeSat to analyze satellite images of 15 beaches across Sumatra, split between the muddy, mangrove-heavy coastline of Dumai in the east and the sandy, grassy shoreline of Padang in the west. in place In areas overgrown with mangrove trees, dense vegetation, or oil palm plantations, measurements were accurate to within 7 meters. In patchy areas, such as grassy areas on beaches or sparsely forested areas, VedgeSat’s accuracy decreased and the margin of error increased to more than 50 meters. However, by manually adjusting the tool, the team found that they could take into account differences in terrain and seawater color, reducing error rates.
“It is very encouraging that VedgeSat can analyze tropical coasts in the same way as the coasts of the countries in which it was developed,” said Dr. Freya Muir from Glasgow, who helped develop VedgeSat and co-authored the paper.
“We hope these results will facilitate the uptake of tools around the world to meet the challenges of a changing planet.”
