Enhancing Chlorophyll Content Estimation With Sentinel-2 Imagery: A Fusion Of Deep Learning And Biophysical Models

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F86652079%3A_____%2F24%3A00601484" target="_blank" >RIV/86652079:_____/24:00601484 - isvavai.cz</a>

Výsledek na webu

<a href="https://ieeexplore.ieee.org/xpl/conhome/10640349/proceeding" target="_blank" >https://ieeexplore.ieee.org/xpl/conhome/10640349/proceeding</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1109/IGARSS53475.2024" target="_blank" >10.1109/IGARSS53475.2024</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Enhancing Chlorophyll Content Estimation With Sentinel-2 Imagery: A Fusion Of Deep Learning And Biophysical Models

Popis výsledku v původním jazyce

Understanding the status of plant traits associated with photosynthesis is essential to gain insights into the carbon dynamics of forest ecosystems and their interconnectedness with global climate change processes. Chlorophyll is a key biochemical component as it is integral to the photosynthetic mechanism. Chlorophyll content (Cab) is therefore widely used to track vegetation dynamics and assess the health status of canopies. In this paper, we present an innovative approach to accurately estimate chlorophyll content at the canopy level by integrating radiative transfer models (RTMs) with deep learning algorithms. Our methodology utilizes a U-Net deep learning model trained with Sentinel-2 (SE2) imagery and the chlorophyll content obtained from resampled SE2 reflectance simulations conducted with the Pro4SAIL model. We evaluated the resulting Cab predictions against leaf chlorophyll content measured in Lanžhot forest, Czechia, during six field campaigns in 2019 and 2020. Our models effectively captured the complexities of forest canopies in Lanžhot forest, where they achieved an RMSE of 8.4 µg/cm2 and MAE of 6.22 µg/cm2. The findings highlight the potential of integrating deep learning algorithms with RTMs and Sentinel-2 time series for chlorophyll content retrieval. The approach not only reduces processing time but also eliminates bottlenecks associated with extrapolating plant trait inversions from RTMs to satellite images. The proposed method significantly enhances the feasibility of integrating RTMs into near real-time monitoring systems. This integration holds promise to the track biochemical effects of biotic (e.g., bark beetle attacks) and abiotic forest disturbances (e.g., drought). By leveraging the synergy of RTMs, deep learning algorithms, and Sentinel2 imagery, this work addresses the challenges of timely and accurate monitoring in dynamic forest ecosystems.

Název v anglickém jazyce

Enhancing Chlorophyll Content Estimation With Sentinel-2 Imagery: A Fusion Of Deep Learning And Biophysical Models

Popis výsledku anglicky

Understanding the status of plant traits associated with photosynthesis is essential to gain insights into the carbon dynamics of forest ecosystems and their interconnectedness with global climate change processes. Chlorophyll is a key biochemical component as it is integral to the photosynthetic mechanism. Chlorophyll content (Cab) is therefore widely used to track vegetation dynamics and assess the health status of canopies. In this paper, we present an innovative approach to accurately estimate chlorophyll content at the canopy level by integrating radiative transfer models (RTMs) with deep learning algorithms. Our methodology utilizes a U-Net deep learning model trained with Sentinel-2 (SE2) imagery and the chlorophyll content obtained from resampled SE2 reflectance simulations conducted with the Pro4SAIL model. We evaluated the resulting Cab predictions against leaf chlorophyll content measured in Lanžhot forest, Czechia, during six field campaigns in 2019 and 2020. Our models effectively captured the complexities of forest canopies in Lanžhot forest, where they achieved an RMSE of 8.4 µg/cm2 and MAE of 6.22 µg/cm2. The findings highlight the potential of integrating deep learning algorithms with RTMs and Sentinel-2 time series for chlorophyll content retrieval. The approach not only reduces processing time but also eliminates bottlenecks associated with extrapolating plant trait inversions from RTMs to satellite images. The proposed method significantly enhances the feasibility of integrating RTMs into near real-time monitoring systems. This integration holds promise to the track biochemical effects of biotic (e.g., bark beetle attacks) and abiotic forest disturbances (e.g., drought). By leveraging the synergy of RTMs, deep learning algorithms, and Sentinel2 imagery, this work addresses the challenges of timely and accurate monitoring in dynamic forest ecosystems.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

20705 - Remote sensing

Projekt

<a href="/cs/project/LM2023048" target="_blank" >LM2023048: Česká infrastruktura sledování uhlíku</a><br>

Návaznosti

I - Institucionalni podpora na dlouhodoby koncepcni rozvoj vyzkumne organizace

Rok uplatnění

2024

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

IEEE International Symposium on Geoscience and Remote Sensing (IGARSS)

ISBN

979-8-3503-6032-5

ISSN

2153-7003

e-ISSN

—

Počet stran výsledku

9400

Strana od-do

4486-4489

Název nakladatele

IEEE Geoscience and Remote Sensing Society (GRSS)

Místo vydání

Athény

Místo konání akce

Athény

Datum konání akce

7. 7. 2024

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

—