Drone-mounted laser scanners have moved from experimental gadgets to core tools of modern forest inventory, but the scientific literature on them is far from evenly spread around the world. A new bibliometric analysis of 66 peer-reviewed documents published between 2015 and 2025 shows that China has become the dominant hub for research on UAV-LiDAR, the pairing of unmanned aerial vehicles with Light Detection and Ranging sensors, while Africa remains almost entirely absent from the field. The study, published in Discover Forests, was conducted by Francis Kamangadazi, Abel Ramoelo and Paxie W. Chirwa, who mined the Scopus and Web of Science databases for English-language papers on drone laser scanning applied to forest structure attributes.
The authors retrieved 107 candidate documents and, after screening and cleaning, retained 66 for analysis. The corpus grew at an average annual scientific production growth rate of 14.87 percent, with a mean document age of 3.8 years and an average of 36.42 citations per document. Journal articles made up the bulk of the dataset, accounting for 60 of the 66 documents, supplemented by two conference papers, two proceedings papers and two reviews. The researchers ran their bibliometric workflows using the open-source Bibliometrix R-package within the RStudio environment, supported by Biblioshiny, a web-based graphical interface that allowed them to map citation networks, co-authorship patterns and thematic clusters across the literature.
Two journals stand out as the primary outlets for this specialized domain. Remote Sensing published 19 of the analyzed articles, making it the most frequent venue, while Forests came second with 7 articles. Ecological Indicators, Remote Sensing of Environment and the International Journal of Applied Earth Observation also contributed meaningfully. On the authorship side, Li X. emerged as the most prolific scholar with 9 publications, followed by Du H. with 6. Xuan J. and Zhang X. have 5 publications each, while six other researchers, including Chen C., Gong Y., Wang D., Wang L., Zhang B. and Zhang Y., each have 4. Notably, the remaining top ten authors are all affiliated with Chinese institutions, a concentration that mirrors the institutional rankings in the study.
Institutionally, Zhejiang A&F University leads with 31 publications, followed by Beijing Forestry University and Northeast Forestry University with 16 each and Nanjing Forestry University with 11. At the country level, China, the United States and Brazil are the pacesetters in publication output, with Norway, Spain, the United Kingdom, Russia and Indonesia also frequently publishing. Cameroon is the sole African country that recorded a research study in the subject area. When it comes to citations, China dominates with 1,273 total citations, followed by the United States with 590 and Norway with 288. Interestingly, Norway has the highest average article citation rate, equal to its total citation count, while the United States averages 118 citations per article, the Netherlands 38, China 28.9, and Denmark 26.
The most cited individual papers reveal the intellectual foundations of the field. Sankey’s 2017 publication tops the list with 401 total citations, cited at 40.1 per year, and evaluated ultra-high-resolution UAV-LiDAR fused with hyperspectral and multispectral datasets to classify individual vegetation species and model three-dimensional canopy architectures. That study found that multi-sensor data fusion yielded an overall accuracy of 88 percent, superior to standalone datasets, particularly for differentiating tree species with overlapping spectral signatures but distinct structural canopy geometries. Puliti’s 2015 paper, the earliest identified work using UAV data in this context, ranks second with 288 citations, cited 24 times per year. It fused three-dimensional structural metrics from photogrammetric drone imagery with ground-truth reference data to fit predictive linear regression models, demonstrating that drone-derived datasets can deliver highly accurate, near-real-time forest inventory parameters even under unfavourable weather conditions.
A third landmark study by Liu and colleagues published the first evaluation of a dedicated UAV-LiDAR configuration for estimating forest structure attributes within plantation forests. The researchers compared the modeling efficacy of plot-level metrics against individual tree-isolated metrics extracted from UAV-LiDAR point clouds using both parametric and non-parametric algorithms. They also ran a sensitivity analysis to assess how laser point density affects structural metric extraction and individual tree segmentation in relation to aboveground biomass modeling. Their results showed that predictive models incorporating combined structural metrics outperformed those restricted purely to plot-level variables, a finding that helped establish the analytical workflows now common across the field.
Keyword analysis reveals where the science is heading. The most recurrent author keywords include aboveground biomass, airborne LiDAR, LiDAR, UAV-LiDAR, remote sensing, UAV, classification and tree species classification. The co-occurrence network partitions the domain into five main thematic clusters, with aboveground biomass at the center of the linkages, followed by the airborne- and UAV-LiDAR network, biomass connected to remote sensing, LiDAR linked to classification and aboveground biomass, and tree species classification linked to point clouds and algorithms. Thematic mapping places UAV-LiDAR, airborne LiDAR and individual trees in the motor themes quadrant, meaning they are both well developed and influential, while clusters around remote sensing, vegetation and forestry, and around aboveground biomass, random forest and Sentinel-2, occupy the niche themes quadrant. A cluster of LiDAR, classification and UAV appears in the emerging or declining themes quadrant, and the authors interpret these terms as pointing to future research opportunities.
The geographic imbalance is stark. The spatial distribution map shows a strong bilateral research collaboration network between China and the United States, with recognizable collaborative links to Spain and Brazil, while no network has been established with Africa. The authors attribute Africa’s limited involvement to limited funding, scarce high-resolution datasets, constrained computational infrastructure and strict aviation policies. Processing millions of discrete LiDAR points requires substantial high-performance computing infrastructure and specialized software to harmonize spatial point clouds with optical imagery. Wide-scale drone integration also faces institutional bottlenecks, including high hardware acquisition costs, a deficit of certified remote pilots, institutional resistance to disruptive technologies, operational human errors, restrictive national airspaces and poorly defined aviation safety regulations. A major technical drawback of UAV-LiDAR systems remains their heavy payload mass, which necessitates larger, more expensive heavy-lift drone platforms and introduces high procurement costs for the sensors themselves, though continuous engineering innovations in miniaturized sensors are projected to bring hardware prices down.
Why does this technology matter so much for forestry? LiDAR instruments scan the Earth’s surface at sampling rates exceeding 150 kilohertz, emitting laser pulses and measuring the time required for the light beam to reach the target and reflect back, which yields dense networks of highly accurate, three-dimensional georeferenced elevation points known as point clouds. When a laser pulse encounters multi-layered targets like forest canopies, multiple return signals are recorded, with the first return typically representing the elevation near the top of the outermost canopy profile and later returns capturing internal branches, trunks or understory vegetation. Analysts use these distinct returns to calculate tree heights and vertical profile metrics. Operating UAV-LiDAR at ultra-low flight altitudes amplifies point densities, achieving concentrations as high as 150 points per square meter in forested areas, which diminishes geometric errors and enhances data extraction from sub-canopy and lower understory vegetation layers. These point clouds allow researchers to extract inventory attributes including diameter at breast height, tree height, stem volume, basal area, forest stand density, canopy cover, aboveground biomass and complex three-dimensional forest structures, metrics essential for assessing forest productivity, ecosystem health, species distribution dynamics and carbon sequestration potential.
The authors argue that the field is undergoing a paradigm shift away from historical, broad macro-level spaceborne assessments toward highly specialized, high-density applications, with precision single-tree modeling and taxonomically explicit tree species classification now established as mature motor themes. Yet those advancements remain highly localized within hyper-productive Western and Asian research corridors. To bridge the regional data divide, the study calls for international consortia to move away from extractive parachute science toward equitable knowledge-transfer frameworks that directly empower African institutions, for future research to validate structural and biomass models within under-studied African biomes such as the Miombo woodlands using locally led field teams, for a transition to open-source and low-computing analytics to circumvent capital and infrastructure deficits in developing nations, and for progressive, standardized aviation frameworks co-designed with civil aviation authorities. The review ultimately confirms the strong potential of UAV-LiDAR for forest structure attribute estimation and serves as an incentive to explore the technology further, particularly in developing nations such as Malawi, where national forest inventories have documented a deforestation and forest degradation rate of approximately 2.8 percent, representing an annual average loss of 250,000 hectares of forest cover.
Subject of Research: Bibliometric analysis of UAV-LiDAR research on forest structure attribute estimation
Article Title: A bibliometric analysis on the use of UAV-LiDAR in forest structure attributes estimation
Article References: Kamangadazi, F., Ramoelo, A., & Chirwa, P. W. (2026). A bibliometric analysis on the use of UAV-LiDAR in forest structure attributes estimation. Discover Forests, 2(1), Article 83. https://doi.org/10.1007/s44415-026-00146-0
Image Credits: AI Generated
DOI: 10.1007/s44415-026-00146-0
Keywords: UAV-LiDAR, forest inventory, bibliometrics, aboveground biomass, tree species classification, remote sensing, point clouds, China, Africa, forest structure, drone forestry, citation analysis
News Source: Alan Morgan. (November 5, 2026). Drone lasers map forests from above, but China dominates the research. Scienmag.
