Skip to main content

ORIGINAL RESEARCH article

Front. Remote Sens.
Sec. Image Analysis and Classification
Volume 6 - 2025 | doi: 10.3389/frsen.2025.1498217
This article is part of the Research Topic Advancements in Agricultural Monitoring with AI Enhanced Remote Sensing Techniques View all 3 articles

AI-Powered Estimation of tree covered area and number of trees over the Mediterranean island of Cyprus

Provisionally accepted
Anna Zenonos Anna Zenonos 1*Sizhuo Li Sizhuo Li 2Martin Brandt Martin Brandt 2Jean Sciare Jean Sciare 1Philippe Ciais Philippe Ciais 3
  • 1 The Cyprus Institute, Nicosia, Cyprus
  • 2 University of Copenhagen, Copenhagen, Capital Region of Denmark, Denmark
  • 3 UMR8212 Laboratoire des Sciences du Climat et de l'Environnement (LSCE), Gif-sur-Yvette, Île-de-France, France

The final, formatted version of the article will be published soon.

    Trees play a crucial role in mitigating climate change by absorbing CO2 and providing biophysical cooling. The European Commission's climate policies underscore the importance of forest monitoring systems to achieve substantial greenhouse gas reductions by 2030. In Cyprus, an EU member state located in the Eastern Mediterranean, and a climate change hot-spot, increasingly impacted by forest fires and more arid conditions, the absence of a comprehensive tree monitoring system hinders effective carbon stock assessment and land-based mitigation strategies. The exact tree population inside and outside forests is currently unknown. Artificial Intelligence is a powerful tool that can enable the development of tree monitoring systems by applying machine learning models to high-resolution image data. This study presents a deep learning neural network model applied to high resolution (10 cm) airborne images collected during the year 2019, to generate segmented tree crowns and the number of individual trees over selected areas of Cyprus, including a large national forest park, a forest park in the capital city, and a small urban area, encompassing a total studied area of 107km2 . The model, previously applied in Denmark and Finland was completely re-tuned using local annotations to account for Cyprus's specific conditions and achieved an overall accuracy of 90% and 93% to estimate the area covered by tree crowns and the number of trees, respectively. The results are regressed against coarser resolution tree cover maps to predict the area covered by tree crowns at a national level. The accuracy of the tree cover maps created by this study is compared to those of existing global tree cover maps, such as the Copernicus products. This work lays the foundation for establishing a tree-level inventory for Cyprus using airborne remote-sensing.

    Keywords: Tree segmentation, deep learning, remote sensing, image analysis, Individual trees

    Received: 18 Sep 2024; Accepted: 06 Jan 2025.

    Copyright: © 2025 Zenonos, Li, Brandt, Sciare and Ciais. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

    * Correspondence: Anna Zenonos, The Cyprus Institute, Nicosia, Cyprus

    Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.