Fusion-Based Approaches and Machine Learning Algorithms for Forest Monitoring: A Systematic Review DOI Open Access

Abdullah Al Saim,

Mohamed H. Aly

Wild, Journal Year: 2025, Volume and Issue: 2(1), P. 7 - 7

Published: March 11, 2025

Multi-source remote sensing fusion and machine learning are effective tools for forest monitoring. This study aimed to analyze various techniques, their application with algorithms, assessment in estimating type aboveground biomass (AGB). A keyword search across Web of Science, Science Direct, Google Scholar yielded 920 articles. After rigorous screening, 72 relevant articles were analyzed. Results showed a growing trend optical radar fusion, notable use hyperspectral images, LiDAR, field measurements fusion-based Machine particularly Random Forest (RF), Support Vector (SVM), K-Nearest Neighbor (KNN), leverage features from fused sources, proper variable selection enhancing accuracy. Standard evaluation metrics include Mean Absolute Error (MAE), Root Squared (RMSE), Overall Accuracy (OA), User’s (UA), Producer’s (PA), confusion matrix, Kappa coefficient. review provides comprehensive overview prevalent data by synthesizing current research highlighting fusion’s potential improve monitoring The underscores the importance spectral, topographic, textural, environmental variables, sensor frequency, key gaps standardized protocols exploration multi-temporal dynamic change

Language: Английский

Estimation of Aboveground Biomass for Different Forest Types Using Data from Sentinel-1, Sentinel-2, ALOS PALSAR-2, and GEDI DOI Open Access
Chu Wang,

Wangfei Zhang,

Yongjie Ji

et al.

Forests, Journal Year: 2024, Volume and Issue: 15(1), P. 215 - 215

Published: Jan. 21, 2024

Forest aboveground biomass (AGB) is integral to the global carbon cycle and climate change study. Local regional AGB mapping crucial for understanding stock dynamics. NASA’s ecosystem dynamics investigation (GEDI) combination of multi-source optical synthetic aperture radar (SAR) datasets have great potential local estimation mapping. In this study, GEDI L4A data ground sample plots worked as true values explore their difference estimating forest using Sentinel-1 (S1), Sentinel-2 (S2), ALOS PALSAR-2 (PALSAR) data, individually in different combinations. The effects types validation were investigated well. S1 S2 performed best with R2 ranging from 0.79 0.84 RMSE 7.97 29.42 Mg/ha, used truth data. While product working reference, range 0.36 0.47 31.41 37.50 Mg/ha. between plot reference shows obvious dependence on types. summary, dataset its SAR better when average less than 150 predictions underperformed across study sites. However, can work source a certain level accuracy.

Language: Английский

Citations

19
Aboveground biomass modeling using simulated Global Ecosystem Dynamics Investigation (GEDI) waveform LiDAR and forest inventories in Amazonian rainforests DOI
Nadeem Fareed, Izaya Numata, Mark A. Cochrane

et al.

Forest Ecology and Management, Journal Year: 2025, Volume and Issue: 578, P. 122491 - 122491

Published: Jan. 5, 2025

Language: Английский

Citations

4
Integrating GEDI, Sentinel-2, and Sentinel-1 imagery for tree crops mapping DOI Creative Commons
Esmaeel Adrah, Jesse P. Wong, He Yin

et al.

Remote Sensing of Environment, Journal Year: 2025, Volume and Issue: 319, P. 114644 - 114644

Published: Feb. 11, 2025

Language: Английский

Citations

2
Development of forest aboveground biomass estimation, its problems and future solutions: A review DOI Creative Commons
Taiyong Ma,

Chao Zhang,

Liping Ji

et al.

Ecological Indicators, Journal Year: 2024, Volume and Issue: 159, P. 111653 - 111653

Published: Feb. 1, 2024

Forest aboveground biomass (AGB) is crucial as it serves a fundamental indicator of the productivity, biodiversity, and carbon storage forest ecosystems. This paper presents targeted literature review advancements in AGB estimation methods. We conducted an extensive published using Web Science, ResearchGate, Semantic Scholar, Google Scholar. Our findings highlight importance accurate studies terrestrial cycle, ecosystem management, climate change. Moreover, contributes valuable ecological knowledge supports effective natural resource management. Unfortunately, during data collection process for estimation, we have identified two critical yet often overlooked issues: (1) reliability manual survey accuracy, (2) impact overlap between ground plots remote sensing pixels on estimation. Drawing existing technologies analysis, propose potentially solution to address these challenges. In conclusion, mapping parameters, such AGB, will remain priority forestry research foreseeable future. To ensure practical applicability findings, our future efforts focus understanding accuracy determining optimal pixels.

Language: Английский

Citations

15
Estimation of above ground biomass in tropical heterogeneous forests in India using GEDI DOI Creative Commons

Indu Indirabai,

Mats Nilsson

Ecological Informatics, Journal Year: 2024, Volume and Issue: 82, P. 102712 - 102712

Published: June 30, 2024

Quantifying above ground biomass (AGB) and its spatial distribution can significantly contribute to monitor carbon stocks as well the storage dynamics in forests. For effective forest monitoring management case of complex tropical Indian forests, there is a need obtain reliable estimates amount sequestration at regional national levels, but estimation quite challenging. The main objective study validate usefulness gridded density (AGBD) (ton/ha) spaceborne LiDAR Global Ecosystem Dynamics Investigation data (GEDI L4B, Version 2) across two heterogeneous forests India, Betul Mudumalai Methodology includes, for each area, linear regression model which predicts AGB from Sentinel-2 MSI was developed using reference comparing it with GEDI AGBD values. Central India had RMSE 13.9 ton/ha, relative = 8.7% R2 0.88, bias −0.28 comparison between modelled 1 km resolution show relatively strong correlation (0.66) no or little bias. It also found that footprint value underestimated compared according model. southern an 29.1 10.8%, 0.79 −0.022. 0.84, field values lies 42.2 ton/ha 238.8 75.9 353.6 ton/ha. results indicates underestimates AGB, used produce product needs be adjusted provide information on balance changes over time type exists test areas.

Language: Английский

Citations

12
Forest aboveground biomass estimation by GEDI and multi-source EO data fusion over Indian forest DOI
Jayantrao Mohite, Suryakant Sawant, Ankur Pandit

et al.

International Journal of Remote Sensing, Journal Year: 2024, Volume and Issue: 45(4), P. 1304 - 1338

Published: Feb. 2, 2024

Monitoring changes in carbon stocks through forest biomass assessment is crucial for cycle studies. However, challenges obtaining timely and reliable ground measurements hinder creation of the spatially continuous maps aboveground density (AGBD). This study proposes an approach generating (AGBD) by combining Global Ecosystem Dynamics Investigation (GEDI) LiDAR-based data with open-access earth observation (EO) data. The key contribution lies systematic evaluation various model configurations to select optimal AGBD generation. considered configurations, including predictor sets, spatial resolution, beam selection, sensitivity thresholds. We used a Random Forest model, trained five-fold cross-validation on 80% total data, estimate Indian region. Model performance was assessed using 20% independent test dataset. Results, Sentinel-1 2 predictors, yielded R2 values 0.55 0.60 RMSE 48.5 56.3 Mg/ha. Incorporating agroclimatic zone attributes improved (R2: 0.59 0.69, RMSE: 42.2 53.3 Mg/ha). selection top 15 which favoured features from Sentinel-2, DEM, attributes, zones, GEDI >0.98, 0.64 46.59 results underscore significance incorporating like agro-climatic zones need considering types shot characteristics. top-performing validated Simdega, Jharkhand 0.74, 39.3 Mg/ha), demonstrating methodological potential this approach. Overall, emphasizes prospects integrating multi-source EO produce (AGB) fusion.

Language: Английский

Citations

10
Forest Aboveground Biomass Estimation Using Multisource Remote Sensing Data and Deep Learning Algorithms: A Case Study over Hangzhou Area in China DOI Creative Commons
Xin Tian, Jiejie Li,

Fanyi Zhang

et al.

Remote Sensing, Journal Year: 2024, Volume and Issue: 16(6), P. 1074 - 1074

Published: March 19, 2024

The accurate estimation of forest aboveground biomass is great significance for management and carbon balance monitoring. Remote sensing instruments have been widely applied in parameters inversion with wide coverage high spatiotemporal resolution. In this paper, the capability different remote-sensed imagery was investigated, including multispectral images (GaoFen-6, Sentinel-2 Landsat-8) various SAR (Synthetic Aperture Radar) data (GaoFen-3, Sentinel-1, ALOS-2), estimation. particular, based on inventory Hangzhou China, Random Forest (RF), Convolutional Neural Network (CNN) Networks Long Short-Term Memory (CNN-LSTM) algorithms were deployed to construct models, respectively. estimate accuracies evaluated under configurations methods. results show that data, ALOS-2 has a higher accuracy than GaoFen-3 Sentinel-1. Moreover, GaoFen-6 slightly worse Landsat-8 optical contrast single source, integrating multisource can effectively enhance accuracy, improvements ranging from 5% 10%. CNN-LSTM generally performs better CNN RF, regardless used. combination provided best case achieve maximum R2 value up 0.74. It found majority values study area 2018 ranged 60 90 Mg/ha, an average 64.20 Mg/ha.

Language: Английский

Citations

9
Accuracy assessment of GEDI terrain elevation, canopy height, and aboveground biomass density estimates in Japanese artificial forests DOI Creative Commons

Hantao Li,

Xiaoxuan Li, Tomomichi Kato

et al.

Science of Remote Sensing, Journal Year: 2024, Volume and Issue: 10, P. 100144 - 100144

Published: June 15, 2024

Global forests face severe challenges owing to climate change, making dynamic and accurate monitoring of forest conditions critically important. Forests in Japan, covering approximately 70% the country's land area, play a vital role yet often overlooked global forestry. Japanese are unique, with 50% comprising artificial forests, predominantly coniferous forests. Despite government's extensive use airborne Light Detecting Ranging (LiDAR) assess conditions, these data need more availability frequency. The Ecosystem Dynamics Investigation (GEDI), first Spaceborne LiDAR explicitly designed for vegetation monitoring, is expected provide significant value high-frequency high-accuracy monitoring. To accuracy GEDI reference were gathered from 53,967,770 trees via Aichi Prefecture, Japan. This was then compared corresponding GEDI-derived terrain elevations, canopy heights (GEDI RH98), aboveground biomass density (AGBD) estimates data. research also explored how different factors influence elevation estimates, including type beam, time acquisition (day or night), beam sensitivity, slope. Additionally, effects various structural parameters, such as height-to-diameter ratio, crown length number on height AGBD, investigated. results showed that demonstrated high across slope rRMSE ranging 2.28% 3.25% RMSE 11.68 m 16.54 m. After geolocation adjustment, comparison derived LiDAR-derived accuracy, exhibiting 22.04%. In contrast, AGBD product moderate 52.79%. findings indicated RH98 influenced by whereas mainly impacted ratio. study provided baseline assessment elevation, RH98, Furthermore, this valuable insights into metrics examining potential factors.

Language: Английский

Citations

9
Forest Aboveground Biomass Estimation and Inventory: Evaluating Remote Sensing-Based Approaches DOI Open Access
Muhammad Nouman Khan, Yumin Tan, Ahmad Ali Gul

et al.

Forests, Journal Year: 2024, Volume and Issue: 15(6), P. 1055 - 1055

Published: June 18, 2024

Remote sensing datasets offer robust approaches for gaining reliable insights into forest ecosystems. Despite numerous studies reviewing aboveground biomass estimation using remote approaches, a comprehensive synthesis of synergetic integration methods to map and estimate AGB is still needed. This article reviews the integrated discusses significant advances in estimating from space- airborne sensors. review covers research articles published during 2015–2023 ascertain recent developments. A total 98 peer-reviewed journal were selected under Preferred Reporting Items Systematic Reviews Meta-Analysis (PRISMA) guidelines. Among scrutinized studies, 54 relevant spaceborne, 22 airborne, datasets. empirical models used, random regression model accounted most (32). The highest number utilizing dataset originated China (24), followed by USA (15). datasets, Sentinel-1 2, Landsat, GEDI, Airborne LiDAR widely employed with parameters that encompassed tree height, canopy cover, vegetation indices. results co-citation analysis also determined be objectives this review. focuses on provides accuracy reliability modeling.

Language: Английский

Citations

9
Statewide Forest Canopy Cover Mapping of Florida Using Synergistic Integration of Spaceborne LiDAR, SAR, and Optical Imagery DOI Creative Commons
Monique Bohora Schlickmann, Inácio Thomaz Bueno, Denis Valle

et al.

Remote Sensing, Journal Year: 2025, Volume and Issue: 17(2), P. 320 - 320

Published: Jan. 17, 2025

Southern U.S. forests are essential for carbon storage and timber production but increasingly impacted by natural disturbances, highlighting the need to understand their dynamics recovery. Canopy cover is a key indicator of forest health resilience. Advances in remote sensing, such as NASA’s GEDI spaceborne LiDAR, enable more precise mapping canopy cover. Although provides accurate data, its limited spatial coverage restricts large-scale assessments. To address this, we combined with Synthetic Aperture Radar (SAR), optical imagery (Sentinel-1 GRD Landsat–Sentinel Harmonized (HLS)) data create comprehensive map Florida. Using random algorithm, our model achieved an R2 0.69, RMSD 0.17, MD 0.001, based on out-of-bag samples internal validation. Geographic coordinates red spectral channel emerged most influential predictors. External validation airborne laser scanning (ALS) across three sites yielded 0.70, 0.29, −0.22, confirming model’s accuracy robustness unseen areas. Statewide analysis showed lower southern versus northern Florida, wetland exhibiting higher than upland sites. This study demonstrates potential integrating multiple sensing datasets produce vegetation maps, supporting management sustainability efforts

Language: Английский

Citations

1