核心概念
本文介紹了一個基於深度學習的紅樹林監測系統,該系統利用 Sentinel-2 衛星數據成功繪製了阿聯酋地區的紅樹林分佈變化,並分析了其背後的驅動因素。
統計資料
阿聯酋的紅樹林面積從 2017 年的 7,080.88 公頃增加到 2024 年的 9,142.21 公頃,總增長率為 29.11%。
從 2017 年到 2024 年,阿聯酋紅樹林的年均增長率約為 3.71%。
阿布達比的紅樹林面積從 2017 年的 5,530.08 公頃增加到 2024 年的 7,385.68 公頃,增長了 33.5%。
模型在紅樹林提取方面取得了優異的性能,生產者精度(召回率)為 0.9196,使用者精度為 0.8917,F1 分數達到 0.9055。
阿聯酋環境署 - 阿布達比 (EAD) 制定了一個雄心勃勃的目標,即到 2030 年種植 1 億棵紅樹林。
引述
“As one of the important countries with mangrove distribution on the Arabian Peninsula, the changes in the mangrove ecosystem in the UAE have attracted much attention.”
“This study aims to use the UNet++ model combined with Sentinel-2 multispectral data to monitor the spatiotemporal changes in mangrove forest of the UAE from 2017 to 2024.”
“The UNet++ model, as an improved deep learning semantic segmentation architecture, has achieved good performance in medical image segmentation and has gradually been applied to remote sensing image analysis in recent years.”
“The series of mangrove protection and restoration plans implemented by the UAE government is a key factor driving the growth of mangrove areas.”