Abstract: The Normalized Difference Vegetation Index (NDVI) derived from polar-orbiting satellites is widely used for vegetation monitoring; however, its temporal continuity is often limited by cloud contamination and fixed revisit cycles. This study investigates the feasibility of using geostationary satellite observations to support NDVI gap filling applications and continuous regional monitoring. Geostationary Ocean Color Imager II (GOCI-II) data were used as input, while Sentinel-2 Multispectral Instrument (MSI) NDVI served as the primary reference dataset. Landsat Operational Land Imager NDVI was additionally employed for independent cross-sensor comparison. A data-driven transformation framework was developed and applied to convert GOCI-II NDVI into MSI-equivalent NDVI while maintaining physically interpretable NDVI values. The transformed NDVI was evaluated through spatial comparisons and pixel-level statistical metrics, including correlation coefficient, mean absolute error, root mean square error, and structural similarity index measure. The results indicate that NDVI transformed from geostationary observations can capture broad spatial patterns and relative variability observed in MSI NDVI, particularly at the field scale. At the same time, reduced contrast and NDVI underestimation are observed, mainly due to spatial resolution differences and sub-pixel heterogeneity. This study emphasizes the potential role of geostationary satellite data as a complementary source for polar-orbiting NDVI products. The findings suggest that integrating geostationary and polar-orbiting satellite observations may contribute to improving NDVI continuity and supporting sustained vegetation monitoring over fixed regions where high temporal resolution is required.

Abstract: The orbital drift of the Fengyun 4B (FY-4B) satellite from 133°E to 105°E in early 2024 significantly altered its viewing geometry over China, providing a unique opportunity to evaluate the impact of satellite positioning on retrieval accuracy. (1) Methods: This study systematically evaluates the performance of FY-4B surface downward shortwave radiation (DSSR) products before and after the drift, using ground radiation observation data from the China Meteorological Administration (CMA) as a reference, including 165 stations. (2) Results: The results demonstrate a substantial improvement in product accuracy post-drift. The correlation coefficient (R) increased from 0.93 to 0.95, while the root mean square error (RMSE) decreased by 11.8% (from 112.46 to 99.24 W/m²). The mean bias error (MBE) shifted from a negligible negative bias to a slight positive bias (2.68 W/m²), indicating reduced systematic deviation. Spatially, the "East-West" accuracy disparity was mitigated, attributed to the reduced viewing zenith angles (VZA) and minimized geometric distortions over western China. While performance over water bodies and urban areas is robust, challenges persist in complex terrains due to 3D topographic effects. (3) Conclusions: Ultimately, the validated high-fidelity radiative records position FY-4B as a reliable data source for solar energy resource assessment, while the demonstrated geometric benefits offer strategic guidance for the orbital deployment of future geostationary constellations.

Abstract: Remote sensing inversion of water quality parameters is a critical interdisciplinary field integrating remote sensing technology, environmental science and water resources management, providing key technical support for precise water resources monitoring and ecological governance. To address the lack of comprehensive systematic reviews in this field, this study conducted a bibliometric-based narrative review, selecting 1058 valid English literatures published during 1995–2025 from the Web of Science Core Collection (WOSCC) and performing in-depth knowledge mapping analysis via CiteSpace software. The results showed that global research in this field has gone through three stages: initial exploration (1997–2009), slow growth (2010–2017) and rapid explosion (2018–2025). China ranks first in publication volume worldwide, with a collaborative research pattern dominated by core institutions including the Chinese Academy of Sciences, Wuhan University and the National Aeronautics and Space Administration (NASA). The core research hotspots focus on multi-source data fusion, AI-driven inversion model optimization, and the research shift from coastal to inland water bodies. Current research faces three key challenges: poor adaptability of multi-source data fusion technologies to water quality monitoring, inadequate integration of geospatial and thematic factors in inversion models and insufficient systematicness of inland water body research. Accordingly, future research should focus on strengthening inland water body studies, advancing remote sensing data fusion methods, and further optimizing water quality inversion models. This study clarifies the field’s development context and research characteristics, providing valuable references for subsequent academic exploration and practical applications in water resources management.

Abstract: Remote sensing has become a cornerstone of data-driven decision-making for monitoring biodiversity and supporting the achievement of the Sustainable Development Goals (SDGs). By providing consistent, spatially explicit observations across scales, Earth observation (EO) technologies enable systematic assessment of environmental change and ecosystem dynamics. Within this context, the Essential Biodiversity Variables (EBV) framework offers a standardised approach to harmonising biodiversity observations from in-situ and remote sensing platforms, thereby enhancing interoperability and the effective use of biodiversity information for conservation and sustainable development. This paper focuses on two EBV classes of particular relevance to EO applications: Ecosystem structure and Species traits. We review recent advances in remote sensing techniques—particularly LiDAR, multispectral, hyperspectral, and radar data—and their capacity to monitor ecosystem vertical structure, ecosystem distribution, habitat suitability, and vegetation traits such as productivity, phenology, leaf area index, chlorophyll content, and functional traits. The integration of EO data with in-situ observations and machine learning approaches is highlighted as a key pathway for improving habitat modelling and biodiversity assessments at regional to continental scales, with direct relevance to SDG 15 (Life on Land). We further discuss current challenges, including data resolution limitations, standardisation, computational demands, and the translation of EO-derived indicators into policy-relevant metrics. Finally, we outline future perspectives, emphasising the role of emerging sensor technologies, artificial intelligence, FAIR data principles, and multi-source data integration in advancing EBV monitoring and strengthening the contribution of remote sensing to sustainable ecosystem management and global biodiversity targets.

Abstract: Robust radar object classification is a challenging task, primarily due to the aspect sensitivity limitation of one-dimensional High-Resolution Range Profile (HRRP) data. To address this, we propose Point-HRRP-Net. This multi-modal framework integrates HRRP with 3D LiDAR point clouds via a Bi-Directional Cross-Attention (Bi-CA) mechanism to enable deep feature interaction. Since paired real-world data is scarce, we constructed a high-fidelity simulation dataset to validate our approach. Experiments conducted under strict angular separation demonstrated that Point-HRRP-Net consistently outperformed single-modality baselines. Our results also verified the effectiveness of Dynamic Graph CNN (DGCNN) for feature extraction and highlighted the high inference speed and the potential of Mamba- based architectures for future efficient designs. Finally, this work validates the feasibility of the proposed approach in simulated environments, establishing a foundation for robust object classification in real-world scenarios.

Abstract: Remote sensing has revolutionized monitoring landscapes that are inaccessible or impractical to survey on the ground. Satellite platforms such as Sentinel-2 enable assessment of ecosystem changes over extensive areas with high temporal frequency, while Unmanned Aerial Systems (UAS) offer flexible, ultra-high-resolution observations ideal for site-specific analysis and sensitive environments. This study compares the performance of Sentinel-2 and Phantom 4 multispectral RTK data for monitoring vegetation dynamics in Mediterranean shrubland ecosystems, focusing on the Normalized Difference Vegetation Index (NDVI). Both platforms produced broadly consistent patterns in seasonal and interannual vegetation dynamics. However, UAS outperformed satellite data in capturing fine-scale heterogeneity, regeneration patches, and subtle disturbance responses, particularly in sparsely vegetated or heterogeneous terrain where satellite metrics may be insensitive. The comparison of NDVI across platforms accounted for standardized processing, harmonization, radiometric and atmospheric correction, and spatial resolution differences. Results show platform selection can be optimized according to monitoring objectives: satellite data are well suited for large-scale, long-term ecosystem monitoring and regional environmental modelling, while UAS data provide critical detail for localized management, early stress detection, and restoration prioritization. A combined approach enhances ecosystem disturbance assessments and resource management by binding the strengths of both wide-area coverage and precise spatial detail.

Abstract: This study explores how demystifying Earth Observation (EO) through co-creation path-16 ways and local language can enhance flood resilience and environmental governance in 17 African informal cities. Using case studies from Maiduguri and Hadejia, Nigeria, the re-18 search employed a transdisciplinary mixed-methods design combining rapid evidence as-19 sessment, surveys, participatory workshops (n = 50 stakeholders) integrating simplified 20 Sentinel-1/2 demonstrations, indigenous knowledge mapping, and pre-/post-engagement 21 surveys. Participants (non-experts) were trained to interpret satellite data in both Hausa 22 and English, linking distant teleconnections with local flood experiences. Findings re-23 vealed significant gains in EO literacy and improvements in interpretive confidence, gen-24 der-inclusive participation, and policy engagement. The use of local learning process en-25 abled participants to translate technical EO concepts into locally meaningful narratives, 26 fostering cognitive empowerment and practical application in flood preparedness and ad-27 vocacy. The study demonstrates that data democratization is not only a matter of open 28 access but also of open understanding. It advances a conceptual model linking Demysti-29 fication, Literacy, Empowerment, Co-Production and Resilience, positioning EO as a so-30 cial technology that bridges scientific and indigenous knowledge systems. The findings 31 contribute to debates on decolonizing environmental science and propose a participatory 32 framework for integrating EO into community-based adaptation, legal accountability, and 33 policy reform across Africa’s rapidly urbanizing landscapes.

Abstract: Terahertz waves are located in the "transition zone" between millimeter waves and infrared light. Terahertz video synthetic aperture radar utilizes the high operating frequency, strong radar cross-section intensity, and high azimuth repetition frequency of terahertz waves to detect and track ground moving targets. The conventional methods for detecting moving targets do not take into account the imaging characteristics of moving targets in terahertz video synthetic aperture radar. The Constant False Alarm Rate (CFAR) detection method is used together with other methods to detect moving targets, resulting in unsatisfactory detection performance. This article proposes a new detection method for single channel slow-moving targets in terahertz video SAR based on shadows and light spots, which extracts the features of the shadow and spot areas of the moving target, and determines the position and direction of the moving target through the identification of the shadow and spot areas. The progressiveness of this method is verified by simulation and experimental tests.

Abstract: Earth observation data has significant potential to support sustainable finance. However, despite the interest and rapidly growing availability of earth observation data, uptake and integration is low within the sector. We explore the barriers experienced by the UK financial sector in using earth observation data for sustainable investment decision-making. We take a reflexive approach to explore the intersection of earth observation technology development, sustainable finance and responsible innovation with the intention of identifying opportunities to share understanding and increase responsible uptake within the sector. From our insights, we set out the stakeholders of earth observation data, their data needs and five challenges to the uptake of earth observation data for sustainable financial decision-making. We develop a baseline of needs across stakeholders and propose the inclusion of responsible innovation principles to support the development of earth observation applications for the sector.

Abstract: Estimating surface soil moisture in semi-arid regions is challenging due to its high spatial and temporal variability, the scarcity of in-situ measurements, and the limitations of optical sensors in the presence of cloud cover and vegetation cover. Synthetic Aperture Radar (SAR) sensors, such as Sentinel-1, overcome these constraints by operating in the microwave domain and providing high-resolution data regardless of atmospheric conditions or daylight availability. This enables the application of inverse semi-empirical models, notably the Hallikainen model for the soil dielectric constant and the Dubois model for backscattering. This study proposes an integrated methodology applied to the municipality of Villaconejos (Madrid, Spain) over the period 2015–2025. The approach was initially calibrated on a pilot plot near Balcón del Tajo using field measurements of soil moisture and soil texture data (sand and clay content) obtained from the SoilGrids platform. Starting from Sentinel-1 VV and VH backscatter coefficients, the combined Hallikainen–Dubois model is inverted through an iterative search over a range of volumetric soil moisture values (0.02–0.45 m* m*) and surface roughness values (0.85–2 cm), selecting the parameter pair that minimises the difference between modelled and observed backscatter. The calibrated methodology is then extrapolated across the entire municipality of Villaconejos using Empirical Bayesian Kriging Regression Prediction (EBK-RP), incorporating topographic covariates (digital elevation model, slope, aspect), hydrological covariates (Topographic Wetness Index, TWI), and vegetation covariates (NDVI). The results include annual and seasonal maps of near-surface volumetric soil moisture (0–5 cm depth) at 10 m resolution and, after a geostatistical downscaling procedure, at 2 m resolution. Additional outputs comprise analyses of temporal variations between wet and dry periods and spatial patterns related to land use and topography. The developed methodology provides continuous, high-resolution, operational, and low-cost soil moisture estimates, representing a valuable tool for water resource management and agro-environmental monitoring in semi-arid regions.

Abstract: This study focuses on the ecological vulnerability and its driving mechanism of the Ebinur Lake Basin. Integrating natural factors such as annual average temperature, annual precipitation and elevation, as well as social factors including GDP and population distribution, it systematically evaluates the ecological vulnerability of the basin from 1994 to 2024 by adopting methods like the SRP model, Analytic Hierarchy Process (AHP) and Geodetector. The results show that the overall scale of ecologically vulnerable areas in the basin has presented a shrinking trend over the past 30 years: the area of severe vulnerability reached a peak of 14,270.31 square kilometers in 2004 and then decreased to 13,242.39 square kilometers; the area of slight vulnerability increased by 60.8%; and the proportion of moderate vulnerability has slightly risen since 2014. Spatially, the vulnerability exhibits significant agglomeration characteristics: severe vulnerable areas are concentrated in the mountainous areas of the basin boundary and the eastern region of Ebinur Lake, while slight vulnerable areas are distributed in woodlands and farmlands of alluvial fans in low mountains and hills. Geodetector analysis shows that, fractional vegetation cover, normalized difference vegetation index and land use type are the dominant factors, natural factors and social factors interact significantly.This study provides a scientific basis for ecological protection and sustainable development of the basin.

Abstract: River discharge is a pivotal metric in hydrological and water resources management. To address limitations in traditional hydrological monitoring stations, such as sparse distribution and high data acquisition costs, this study focuses on the Fuyang River LHK hydrological station in Handan City, Hebei Province, China, and proposes a synergistic estimation method for river discharge using multi-source remote sensing data. The approach first extracts river water bodies from Sentinel-1 SAR imagery and Sentinel-2 optical imagery via EN-OTSU and MNDWI-OTSU algorithms, respectively. Subsequently, river width is calculated using the water area-to-length ratio method to reduce errors caused by edge effects. Finally, a power-law discharge estimation model is developed by fitting river width to discharge data. For water body extraction, the Sentinel-2 MNDWI-OTSU method achieves the highest accuracy (overall accuracy: 95.31%, Kappa coefficient: 0.90), followed by the Sentinel-1 EN-OTSU method (overall accuracy: 92.55%, Kappa coefficient: 0.89). For discharge estimation, both data sources exhibit robust inversion performance, with the Sentinel-1-based model showing superior error stability (NSE=0.83, R²=0.83, RRMSE=0.24) and the Sentinel-2-based model marginally better theoretical fit (NSE=0.84, R²=0.84, RRMSE=0.26). Compared with traditional in situ measurements and single-sensor approaches, this method enables a shift from point-based to basin-wide dynamic monitoring, resolving data scarcity in ungauged regions; it integrates the high boundary delineation precision of optical remote sensing with the all-weather penetration of radar, effectively countering interruptions from cloudy and rainy conditions; and it reduces reliance on ground infrastructure, providing a cost-effective, reliable framework for river monitoring and informed water resource allocation.

Abstract: Video synthetic aperture radar (SAR) enables observation of moving targets by leveraging temporal information across successive frames. In particular, dynamic shadows in video SAR image sequences provide critical cues for detecting moving objects whose energy is smeared or Doppler-shifted. To achieve high-resolution imaging at a high frame rate for effective dynamic scene monitoring, video SAR systems typically operate at extremely high frequencies or even in the terahertz band, rather than the microwave band. However, terahertz video SAR suffers from significant signal attenuation due to atmospheric absorption. We present a deep learning framework for high-frame-rate and high-resolution imaging with microwave video SAR system. In this framework, the problem of microwave video SAR imaging is formulated as an image super-resolution reconstruction task for low-resolution yet high-frame-rate image sequences from microwave video SAR. We develop a simple yet effective image super-resolution reconstruction network that is completely built upon convolutional neural networks. The designed network takes a low-resolution image sequence and the corresponding high-resolution image with blurred shadows as input, and then produces a high-resolution image sequence where shadows are clearly visible. Furthermore, the network is trained in a self-supervised manner and thus does not require desired high-resolution image sequences as ground truth, which is appealing to practical applications. Processing results of real data from two different video SAR systems have shown good performance of the proposed approach with convincing generalization ability.

Abstract: Cloud detection is an important procedure for the processing of remote sensing images. A cloud detection scheme driven by the spectral and the temporal features is presented in this paper, where an unsupervised hierarchy clustering approach is proposed for large scale image segmentation. The potential cloudy pixels are identified by means of the spectral matching, in which the spectral data of the clustering centers are compared to the patterns in the spectral dataset of ground covers. The matched pixels are regarded as cloudless pixels, whose category can be recognized accordingly. In contrast, the bright temperatures corresponding to the unmatched pixels are used to exclude the interference of the occasional hotspots, enabling the final cloud detection result. Landsat 8, Sentinel-2, and MODIS satellite data are used in the validation to demonstrate the precision and stability of the proposed scheme for the data at different spatial resolutions.

Abstract: The exponential rise of microsatellite constellations offers unprecedented temporal resolution for urban monitoring. However, ensuring the radiometric integrity of these sensors over heterogeneous built environments remains a critical challenge due to low signal-to-noise ratios and spectral uncertainties. Traditional vicarious calibration relies on homogeneous pseudo-invariant calibration sites (PICS) in deserts, which fail to represent the spectral complexity and adjacency effects of urban landscapes. This study presents a novel triple-platform validation framework integrating ground (Hyperspectral), UAV (Multispectral), and satellite (Sentinel-2) data to bridge the "Point-to-Pixel" scale gap. We introduce a physics-informed "Double Calibration" protocol—combining the empirical line method with spectral response function convolution—and a block kriging spatial upscaling technique to mathematically model intra-pixel heterogeneity. Results from a 2025 campaign in a complex urban environment (Cheongju, Republic of Korea) demonstrate that simple point-averaging introduces significant representation errors (R^2\approx0.46 with time lag). In contrast, our UAV-based block kriging approach recovered high correlations even with a 1-day time lag and dramatically improved the coefficient of determination (R2) under simultaneous acquisition conditions: from 0.68 to 0.92 in the blue band and to 0.96 in the NIR band. Furthermore, quantitative spatial analysis identified artificial grass as the most stable "Urban PICS" (\sigma\approx0.020), whereas Asphalt exhibited unexpected high spatial heterogeneity (\sigma>\ 0.09) due to surface aging, challenging conventional assumptions. This framework establishes a rigorous, scalable standard for validating "New Space" data products in complex urban domains.

Abstract:

Unmanned Aerial Vehicles (UAVs) are essential tools for high-resolution urban remote sensing; however, maximizing their operational efficiency is often hindered by the Size, Weight, and Power (SWaP) constraints inherent to aerial platforms. High-end sensors (e.g., LiDAR) provide dense data but reduce flight endurance and require extensive post-processing, delaying actionable intelligence. To address the challenge of maximizing data utility through cost-effective means, this study evaluates an adaptive multi-modal monitoring framework utilizing high-resolution RGB imagery. Using a DJI Matrice 300 RTK, we assessed the performance of RGB-based advanced AI architectures across varying urban density zones. We stress-tested End-to-End Deep Learning models (Mask R-CNN, YOLOv8-seg) and a Hybrid approach (U-Net++ fused with RGB-derived Canopy Height Models) to determine their viability for replacing active sensors in precision analysis. Results indicate that the RGB-based Hybrid model achieved superior Semantic IoU (0.551), successfully demonstrating that optical imagery combined with deep learning can substitute for heavy active sensors in area-based estimation tasks. Crucially for autonomous UAV operations, YOLOv8-seg achieved inference speeds of 3.89 seconds per tile, approximately 1.86 times faster than Mask R-CNN, validating its suitability for onboard inference on embedded systems. This study establishes a protocol for high-precision analysis using standard RGB sensors, offering a strategic pathway for deploying scalable, consumer-grade UAV fleets in complex urban environments.

Abstract: Due to hardware limitations of Thermal Infrared (TIR) cameras, TIR images captured by Unmanned Aerial Vehicles (UAVs) suffer from Low Resolution (LR) and blurred textures. Improving the spatial resolution of TIR images is of great significance for subsequent applications. Existing image Super-Resolution (SR) methods rely on High-Resolution (HR) ground truth for supervised training, resulting in poor general-ization ability. They also lack constraints on the temperature information of TIR imag-es, failing to maintain the consistency of temperature information reconstruction. To address these two issues, this paper proposes a UAV TIR image SR method based on diffusion models and cross-modal texture transfer, which introduces HR information from Visible (VIS) images into TIR images. Firstly, a Multi-Stage Decomposition Latent Low-Rank Representation (MS-DLatLRR) method is adopted to extract multi-scale de-tailed textures from VIS images. Secondly, prior information of object thermal radia-tion is introduced, and combined with the segmentation map of VIS images, a guided coefficient map for VIS multi-scale detailed texture transfer is constructed to provide constraints for temperature consistency during the cross-modal texture transfer pro-cess. Finally, the multi-scale detailed textures and the guided coefficient map are in-troduced into a diffusion model (MP-DDNM) for SR processing of TIR images. Exper-imental results show that compared with existing methods, the proposed method im-proves the resolution of UAV TIR images while maintaining the consistency of tem-perature information as much as possible.

Abstract: Few-shot object detection (FSOD) in high-resolution remote sensing (RS) imagery remains challenging due to scarce annotations, large intra-class variability, and high visual similarity between categories, which together limit the generalization ability of convolutional neural network (CNN)-based detectors. To address this issue, we explore leveraging large vision-language models (LVLMs) for FSOD in RS. We propose a two-stage, parameter-efficient fine-tuning framework with hierarchical prompting that adapts Qwen3-VL for object detection. In the first stage, low-rank adaptation (LoRA) modules are inserted into the vision and text encoders and trained jointly with a Detection Transformer (DETR)-style detection head on fully annotated base classes under three-level hierarchical prompts. In the second stage, the vision LoRA parameters are frozen, the text encoder is updated using K-shot novel-class samples, and the detection head is partially frozen, with selected components refined using the same three-level hierarchical prompting scheme. To preserve base-class performance and reduce class confusion, we further introduce knowledge distillation and semantic consistency losses. Experiments on the DIOR and NWPU VHR-10.v2 datasets show that the proposed method consistently improves novel-class performance while maintaining competitive base-class accuracy and surpasses existing baselines, demonstrating the effectiveness of integrating hierarchical semantic reasoning into LVLM-based FSOD for RS imagery.

Abstract: Intensifying urbanisation in the Arctic, particularly in spatially constrained coastal and island cities, requires reliable information on long-term land use/land cover (LULC) change to assess environmental impacts and support urban planning. However, multi-decadal, high-resolution LULC datasets for Arctic cities remain limited. In this study, we quantify LULC change on Tromsøya (Tromsø, Norway) from 1984 to 2024 using multispectral satellite imagery based on Landsat and PlanetScope, complemented by LiDAR-derived canopy height models (CHM) and building footprints. We mapped LULC change trajectories and examined how these shifts relate to district-level population redistribution using gridded population data. The integration of a LiDAR-derived CHM was found to substantially improve the accuracy of Landsat-based LULC mapping and to represent the dominant source of classification gains, particularly for spectrally similar urban classes such as residential areas, roads, and other paved surfaces. Landsat augmented with CHM was shown to achieve practical equivalence to PlanetScope when the latter was modelled using spectral features only, supporting the feasibility of scalable and cost-effective long-term monitoring of urbanisation in Arctic cities. Based on the best-performing Landsat configuration, the proportions of artificial and green surfaces were estimated, indicating that approximately 20% of green areas were transformed into artificial classes. Spatially, population growth was concentrated in a small number of districts and broadly coincided with hotspots of green to artificial conversion The workflow provides a reproducible basis for long-term, district-scale LULC monitoring in small Arctic cities where data constraints limit consistent use of high-resolution image.

Abstract: Tropical forests are critical for global climate, biodiversity conservation, and supporting local livelihoods, yet they remain highly vulnerable to human-induced pressures (deforestation and degradation) and climate change impacts (diseases, fires, and drought). The overarching aim of this review is to assess how artificial intelligence (AI) and machine learning (ML) are transforming remote sensing-based monitoring of tropical forests, with a focus on their potential to enhance the detection and estimation of forest change and support tropical forest-related climate policy frameworks. The strengths of this review lie in its comprehensive synthesis of technical, institutional, and governance dimensions, achieved by systematically analyzing evidence from operational forest monitoring platforms and peer-reviewed literature (2010–2025). Using structured search and qualitative analysis, the review evaluates advances in AI/ML applications, identifies technical and institutional barriers, highlights emerging solutions, and provides practical, policy-relevant recommendations. This review identifies critical gaps and proposes a roadmap for scaling AI/ML for tropical forest monitoring. It finds that AI/ML tools, particularly supervised and unsupervised classifiers, deep learning models, time-series analytics, and multi-sensor data-fusion approaches, have become central to advancing remote sensing— enhancing accuracy, automation, and scalability for monitoring deforestation, forest degradation, biomass change, and forest dynamics. However, effective adoption of these technologies still faces persistent barriers—such as limited access to high-quality training data, reliance on proprietary platforms, technical capacity gaps, and unresolved ethical and governance challenges. The review concludes that overcoming these barriers through open training datasets, platform-agnostic infrastructures, capacity building, and inclusive governance is essential for scaling robust, transparent, and locally owned AI-enabled forest monitoring systems. Advances in AI/ML in remote sensing will support climate mitigation, biodiversity conservation, and equitable decision-making in tropical forest countries.