Nyström-based spectral clustering using airborne LiDAR point cloud data for individual tree segmentation

Reliable information concerning stand volume is fundamental to making strategic decisions in sustainable forest management. A variety of remotely sensed data and different inventory methods have been used for the estimation of forest biometric parameters. Particularly, airborne laser scanning (ALS) point clouds are widely used for the estimation of stand volume and forest biomass using an area-based approach (ABA) framework. This method relies on the reference measurements of field plots with the necessary prerequisite of a precise co-registration between ground reference plots and the corresponding ALS samples. In this research, the allometric area-based approach (AABA) is proposed in the context of stand volume estimation of Scots pine (Pinus sylvestris L.) stands. The proposed method does not require detailed information about the coordinates of the field plots. We applied Polish National Forest Inventory data from 9400 circular field plots (400 m2) to develop a plot level stand volume allometric model using two independent variables: top height (TH) and relative spacing index (RSI). The model was developed using the multiple linear regression method with a log–log transformation of variables. The hypothesis was that, the field measurements of TH and RSI could be replaced with corresponding ALS-derived metrics. It was assumed that TH could be represented by the maximum height of the ALS point cloud, while RSI can be calculated based on the number of tree crowns delineated within the ALS-derived canopy height model. Performance of the developed AABA model was compared with the semi-empirical ABASE (with two predictors: TH and RSI) and empirical ABAE (several point cloud metrics as predictors). The models were validated at the plot level using 315 forest management inventory plots (400 m2) and at the stand level using the complete field measurements from 42 Scots pine dominated forest stands in the Milicz forest district (Poland). The AABA model showed a comparable accuracy to the traditional ABA models with relatively high accuracy at the plot (relative root mean square error (RMSE) = 22.8 per cent; R2 = 0.63) and stand levels (RMSE = 17.8 per cent, R2 = 0.65). The proposed novel approach reduces time- and cost-consuming field work required for the classic ABA method, without a significant reduction in the accuracy of stand volume estimations. The AABA is potentially applicable in the context of forest management inventory without the necessity for field measurements at local scale. The transportability of the approach to other species and more complex stands needs to be explored in future studies.

Tree height quantification using very high resolution imagery acquired from an unmanned aerial vehicle (UAV) and automatic 3D photo-reconstruction methods

Surface fuel load (SFL) constitutes one of the most significant fuel components and is used as an input variable in most fire behavior prediction systems. The aim of the present study was to investigate the potential of discrete-return multispectral Light Detection and Ranging (LiDAR) data to reliably predict SFL in a coniferous forest characterized by dense overstory and complex terrain. In particular, a linear regression analysis workflow was employed with the separate and combined use of LiDAR-derived structural and pulse intensity information for the load estimation of the total surface fuels and individual surface fuel types. Following a leave-one-out cross-validation (LOOCV) approach, the models developed from the different sets of predictor variables were compared in terms of their estimation accuracy. LOOCV indicated that the predictive models produced by the combined use of structural and intensity metrics significantly outperformed the models constructed with the individual sets of metrics, exhibiting an explained variance (R2) between 0.59 and 0.71 (relative Root Mean Square Error (RMSE) 19.3–37.6%). Overall, the results of this research showcase that both structural and intensity variables provided by multispectral LiDAR data are significant for surface fuel load estimation and can successfully contribute to effective pre-fire management, including fire risk assessment and behavior prediction in case of a fire event.

ABSTRACTThe normalized difference vegetation index (NDVI) is a commonly used index for monitoring crop growth status. Previous studies have shown that the leaf area index (LAI) estimation based on NDVI is limited by saturation that occurs under conditions of relatively dense canopies (LAI > 2 m2 m–2). To reduce the saturation effect, we suggested new spectral indices through the spectral indices approach. The results suggested that the two-band normalized difference spectral index (NDSI = ((ρ940 – ρ730) /(ρ940 + ρ730))) resulted from the two-band spectral indices approach and the three-band modified normalized difference spectral index (mNDSI = ((ρ940 – 0.8 × ρ950) – ρ730) /((ρ940 – 0.8 × ρ950) + ρ730)) resulted from the three-band spectral indices approach, and they were able to mitigate saturation and improve the LAI prediction with a determination coefficient (R2) of 0.77 and 0.78, respectively. In the validation based on data from independent experiments, these new indices exhibited an accuracy with relative root mean square error (RRMSE) lower than 23.38% and bias higher than –0.40. These accuracies were significantly higher than those obtained with some existing indices with good performance in LAI estimation, such as the enhanced vegetation index (EVI) (RRMSE = 30.19%, bias = –0.34) and the modified triangular vegetation index 2 (MTVI2) (RRMSE = 29.30%, bias = –0.28), and the indices with the ability to mitigate the saturation, such as the wide dynamic range vegetation index (WDRVI) (RRMSE = 31.37%, bias = –0.54), the red-edge wide dynamic range vegetation index (red-edge WDRVI) (RRMSE = 26.34%, bias = –0.54), and the normalized difference red-edge index (NDRE) (RRMSE = 28.41%, bias = –0.56). Additionally, these new indices were more sensitive under moderate to high LAI conditions (between 2 and 8 m2 m–2). Between these two new developed spectral indices, there was no significant difference in the accuracy and sensitivity assessments. Considering the index structure and convenience in application, we demonstrated that the two-band spectral index NDSI((ρ940 – ρ730) /(ρ940 + ρ730)) is efficient in mitigating saturation and has considerable potential for estimating the LAI of canopies throughout the entire growing season of wheat (Triticum aestivum L.), whereas the three-band spectral index contributes lesser in the saturation mitigation provided the red-edge band has been contained.

Accurate extraction of tree parameters is vital for the calculation of high-quality forest volume or biomass. The Light Detection and Ranging (Lidar) technology with its ability to acquire three-dimensional forest stand structures is an important data source for extracting tree parameters. However, complex forest stand structures due to dense canopy and abundant understory vegetation can seriously affect the extraction performance of tree parameters. This study selected Chinese fir plantations to examine how different forest stand structure conditions affected tree parameter extraction based on the integration of handheld laser scanning (HLS) and airborne laser scanning (ALS) data. This study also explored the ability of developing models using extracted tree height and crown width, or directly using the calculated variables from point clouds to estimate the diameter at breast height (DBH) and stem volume. The methods mainly consisted of two parts: (1) Selecting appropriate height to identify seed points for tree segmentation, and subsequently extracting tree parameters; (2) Taking tree height, crown width, and tree-level point cloud metrics as independent variables to establish regression and random forest models for estimating tree DBHs and stem volumes. Results showed that (1) tree DBHs were accurately extracted with relative root mean square error (RMSEr) of 7.1%−10.5% and stem volume with an RMSEr of 12.3%−16.8% under different understory conditions; (2) As stand structure became complex, direct extraction of tree DBHs from HLS data became a challenging task. However, developing a DBH estimation model using tree height and crown width proved feasible; (3) In the complex-understory condition, the volume estimation model utilizing point cloud variables achieved good performance with RMSEr of 28.0%. This research provides new insights for extracting tree parameters using HLS and ALS data, offering a potential replacement for field measurements of tree DBH, height, and stem volume for Chinese fir plantations.

Light detection and ranging (LiDAR) is an active remote sensing technology that can be used to directly acquire three-dimensional forest information, and it has great potential to replace traditional time-consuming field surveys. We aimed to determine the accuracy with which enhanced airborne LiDAR with high density (mean, 47 points m−2) can replicate detailed vertical foliage arrangement and three forest structural indices, i.e. number of layers (NSTRAT), sum of the percent vegetation cover (PCVC), and foliage height diversity (FHD), all of which are useful for characterizing avian habitats. Thirteen 40-m-long lines were set in an urban forest area, and LiDAR-based and field-based cross-sectional profiles with 1 × 1-m meshes were compared. In addition, we performed random LiDAR point thinning to determine whether the replications were possible with lower point densities. The 1 × 1-m mesh vertical foliage arrangements were replicated with high accuracy by using LiDAR (mean producer's accuracy, 0.756; mean user's accuracy, 0.810) before the point reduction. All the three indices were also replicated with high accuracy (<0.12 of the relative root mean square error for the 1:1 relationship). Even when the mean point density was reduced to 2.4 points m−2, the three indices could be quantitatively estimated with high accuracy by using simple linear regression analyses (>0.8 in R 2). For NSTRAT and FHD, numerical replication was possible with high accuracy (<0.2 and <0.1, respectively, of the relative root mean square error for the 1:1 relationship), even when the mean point density was reduced to 2.4 points m−2. In conclusion, enhanced high-density LiDAR seems to be highly suitable as a substitute for field surveys for delineating detailed vertical forest structure in order to characterize forest avian habitats. Moreover, lower-density LiDAR data could replicate the forest structural indices, suggesting its useful low-cost application in quantitative avian habitat models.

BackgroundThe increasing availability of remotely sensed data has recently challenged the traditional way of performing forest inventories, and induced an interest in model-based inference. Like traditional design-based inference, model-based inference allows for regional estimates of totals and means, but in addition for wall-to-wall mapping of forest characteristics. Recently Light Detection and Ranging (LiDAR)-based maps of forest attributes have been developed in many countries and been well received by users due to their accurate spatial representation of forest resources. However, the correspondence between such mapping and model-based inference is seldom appreciated. In this study we applied hierarchical model-based inference to produce aboveground biomass maps as well as maps of the corresponding prediction uncertainties with the same spatial resolution. Further, an estimator of mean biomass at regional level, and its uncertainty, was developed to demonstrate how mapping and regional level assessment can be combined within the framework of model-based inference.ResultsThrough a new version of hierarchical model-based estimation, allowing models to be nonlinear, we accounted for uncertainties in both the individual tree-level biomass models and the models linking plot level biomass predictions with LiDAR metrics. In a 5005 km2 large study area in south-central Sweden the predicted aboveground biomass at the level of 18 m ×18 m map units was found to range between 9 and 447 Mg ·ha−1. The corresponding root mean square errors ranged between 10 and 162 Mg ·ha−1. For the entire study region, the mean aboveground biomass was 55 Mg ·ha−1 and the corresponding relative root mean square error 8%. At this level 75% of the mean square error was due to the uncertainty associated with tree-level models.ConclusionsThrough the proposed method it is possible to link mapping and estimation within the framework of model-based inference. Uncertainties in both tree-level biomass models and models linking plot level biomass with LiDAR data are accounted for, both for the uncertainty maps and the overall estimates. The development of hierarchical model-based inference to handle nonlinear models was an important prerequisite for the study.

Abstract. This paper examines various geometric shapes to determine the most suitable one for calculating tree crown area in field measurements. The study was conducted in an Eldarica pine plantation forest, which was digitally mapped using RGB images captured by a Phantom 4 Pro drone. Tree crowns were manually digitized (MD) from these images to serve as reference data. Field measurements, including the large and small crown diameters, were collected to evaluate crown areas derived from different geometric shapes. The geometric shapes considered were: Oval with Both Diameters (OBD), Circle with Small Diameter (CSD), Circle with Large Diameter (CLD), and Circle with Mean Diameters (CMD). Three analyses were performed to assess the results: correlation analysis (R2), relative root mean square error (RRMSE), and shape analysis, which included overestimation (OverID) and underestimation (UnderID) indices. The results revealed that the choice of geometric shape significantly impacts the accuracy of crown area calculations. The OBD model based on the outer boundary diameter yielded the best results with RRMSE = 0.29, R2 = 0.84, OverID = 0.18, and UnderID = 0.23, followed closely by the CMD method. In contrast, the CSD and CLD models performed less effectively, with RRMSE = 0.52, R2 = 0.42, OverID = 0.11, UnderID = 0.35 (CSD), and RRMSE = 0.59, R2 = 0.37, OverID = 0.46, UnderID = 0.22 (CLD). These differences in performance are likely due to the inclusion of empty spaces within the crown area in some models. However, the findings of this study are not universally applicable to all tree crown area calculations. The geometric shape used for crown area estimation must align with the structural characteristics of the forest and the specific geometry of the tree species under consideration.

Nowadays, forest inventories are frequently carried out using a combination of field measurements and remote sensing data, often acquired with light detection and ranging (LiDAR) sensors. Several studies have investigated how three-dimensional laser scanning point clouds from different platforms can be used to acquire information traditionally collected with forest instruments, such as hypsometers and callipers to detect single-tree attributes like tree height and diameter at the breast height. The present study has tested the performances of the ZEB1 instrument, a type of hand-held mobile laser scanner, for single-tree attributes estimation in pure Castanea sativa Mill. stands cultivated for fruit production in Central Italy. In particular, the influence of walking scan path density on single-tree attributes estimation (number of trees, tree position, diameter at breast height, tree height, and crown base height) was investigated to test the efficiency of field measures. The point clouds were acquired by walking along straight lines drawn with different spacing: 10 and 15 m apart. A single-tree scan approach, which included walking with the instrument around each tree, was used as reference data. In order to evaluate the efficiency of the survey, the influence of the walking scan path was discussed in relation to the accuracy of single-tree attributes estimation, as well as the time and cost needed for data acquisition, pre-processing, and analysis. Our results show that the 10 m scan path provided the best results, with an omission error of 6%; the assessment of single-tree attributes was successful, with values of the coefficient of determination and the relative root mean square error similar to other studies. The 10 m scan path has also proved to decrease the costs by about €14 for data pre-processing, and a saving of time for data acquisition and data analysis of about 37 min compared to the reference data.

Individual tree (IT) segmentation is crucial for forest management, supporting forest inventory, biomass monitoring or tree competition analysis. Light detection and ranging (LiDAR) is a prominent technology in this context, outperforming competing technologies. Aerial laser scanning (ALS) is frequently used for forest documentation, showing good point densities at the tree-top surface. Even though under-canopy data collection is possible with multi-echo ALS, the number of points for regions near the ground in leafy forests drops drastically, and, as a result, terrestrial laser scanners (TLS) may be required to obtain reliable information about tree trunks or under-growth features. In this work, an IT extraction method for terrestrial backpack LiDAR data is presented. The method is based on DBSCAN clustering and cylinder voxelization of the volume, showing a high detection rate (∼90%) for tree locations obtained from point clouds, and low commission and submission errors (accuracy over ). The method includes a sensibility assessment to calculate the optimal input parameters and adapt the workflow to real-world data. This approach shows that forest management can benefit from IT segmentation, using a handheld TLS to improve data collection productivity.

Abstract. A dual-lidar system comprising two WindCube v2.1 Doppler lidars – one oriented horizontally at a 45° angle relative to the other – was deployed to estimate atmospheric turbulence using the variance method. This approach derives second-order velocity statistics directly from line-of-sight (LOS) velocity variances, enabling the reconstruction of the three-dimensional velocity variance components without retrieving instantaneous wind vectors. Its performance is evaluated against the traditional method commonly used in the wind energy sector, which reconstructs instantaneous velocity components from single-lidar LOS measurements prior to computing turbulence statistics. Both methods are assessed at a single measurement height using a 30 d collocated dataset, with sonic anemometer measurements as reference and classification by atmospheric stability. The analysis focuses on turbulence intensity (TI), and performance is quantified using the mean relative bias error (MRBE) and relative root mean square error (RRMSE), following Det Norske Veritas (DNV) load-validation acceptance criteria (MRBE within ±5 % and RRMSE ≤ 15 %). Results show that the variance method demonstrates improved agreement with the reference measurements across nearly all wind-speed bins. For along-wind TI, MRBE values range between 0 % and −5 %, with RRMSE remaining below 15 % over the full wind-speed range, fully satisfying the DNV criteria. For cross-wind TI, the MRBE criterion is met, while the RRMSE threshold is exceeded. In contrast, the traditional method meets both DNV criteria for along-wind TI only at wind speeds above approximately 7 m s−1 and fails to satisfy either criterion for cross-wind TI across the investigated range. Overall, the results demonstrate that the variance method provides more robust and DNV-compliant TI estimates than the traditional reconstruction approach, particularly for along-wind turbulence relevant to wind turbine load validation.

Soil pH is a key factor that controls soil nutrient availability, soil microbial activities, and crop growth and development. However, studies on the soil pH variations of cultivated lands in different horizons at the regional scale remain limited. In this work, 348 soil samples were collected from three soil horizons (A, B, and C) at 120 sites over the hilly region of Chongqing, southwestern China. Six topographic indicators, four climate parameters, and parent material were considered. Classification and regression trees (CARTs) were applied to investigate the relationships between soil pH and the variables in the A, B, and C horizons. Model performances were evaluated by root mean square error (RMSE), relative root mean square error (RRMSE), and coefficient of determination (R2). Results showed that soil pH increased obviously from the A to C horizons. Soil pH was predicted well by the forcing factors with the CART models in all horizons. RMSE, RRMSE, and R2 varied between 0.37 and 0.435, between 5.93 and 7.23%, and between 0.71 and 0.80, respectively. The relative importance of the studied variables to soil pH differed with the horizons. Annual temperature range (ATR), terrain wetness index (TWI), and Melton ruggedness number were critical factors that controlled soil pH variability in the A horizon. Parent material, precipitation of warmest quarter (PWQ), ATR, and TWI were important variables in the B horizon. Parent material, PWQ, ATR, and precipitation were key factors in the C horizon. The results are expected to provide valuable information for designing appropriate measurements for agricultural practices and preventing soil acidification.

Soil water content (SWC) is an important factor restricting crop growth and yield in cropland ecosystems. The observation and simulation of soil moisture contribute greatly to improving water-use efficiency and crop yield. This study was conducted at the Shandong Yucheng Agro-ecosystem National Observation and Research Station in the North China Plain. The study period was across the winter wheat (Triticum aestivum L.) growth stages from 2017 to 2019. A cosmic-ray neutron probe was used to monitor the continuous daily SWC. Furthermore, the crop leaf area index (LAI), yield, and aboveground biomass of winter wheat were determined. The root zone quality model 2 (RZWQM2) was used to simulate and validate the SWC, crop LAI, yield, and aboveground biomass. The results showed that the simulation errors of SWC were minute across the wheat growth stages and mature stages in 2017–2019. The root mean square error (RMSE) and relative root mean square error (RRMSE) of the SWC simulation at the jointing stage of winter wheat were 0.0296 and 0.1605 in 2017–2018, and 0.0265 and 0.1480 in 2018–2019, respectively. During the rain-affected days, the RMSE (0.0253) and RRMSE (0.0980) for 2017–2018 were significantly lower than those of 2018–2019 (0.0301 and 0.1458, respectively), indicating that rain events decreased the model accuracy in the dry years compared to the wet years. The simulated LAIs were significantly higher than the measured values. The simulated yield value of winter wheat was 5.61% lower and 3.92% higher than the measured yield in 2017–2018 and in 2018–2019, respectively. The simulated value of aboveground biomass was significantly (45.48%) lower than the measured value in 2017–2018. This study showed that, compared with the dry and cold wheat growth period of 2018–2019, the higher precipitation and temperature in 2017–2018 led to a poorer simulation of SWC and crop-growth components. This study indicated that annual abnormal rainfall and temperature had a significant influence on the simulation of SWC and wheat growth, especially under intensive climate-change stress conditions.

In this study, we developed a model for photosynthetic production in double cropping rice by integrating the advantages in current crop models (including the models of canopy structure, canopy light distribution, canopy photosynthesis and dry matter production). The canopy light distribution and dry matter accumulation were preliminarily validated with independent field experiment datasets. The distribution of direct radiation both on a level surface and on the leaf surface within canopy, the canopy daily photosynthate and its characteristics with varying leaf area index for three typical plant types (erect both upper and lower, upper erect and lower flat, and flat both upper and lower) were quantitatively analyzed by the model. The results indicated that there was a good agreement between the simulated and observed values. The root mean square error (RMSE), relative root mean square error (RRMSE) and correlation coefficient (r) of prediction of canopy light distribution in double cropping rice were 12.01 J ·m-2·s-1, 8.2% and 0.9929, respectively. Meanwhile, the RMSE, RRMSE and r of prediction of dry matter accumulation were 0.83 t·hm-2, 14.6% and 0.9772, respectively. It was indicated that the model had a performance. The upper erect and lower flat plant type had highest canopy daily photosynthate due to higher incident sun light received on the leaf surface, leaf photosynthetic efficiency and leaf area index.

Resistance value (R-value) is one of the basic subgrade stiffness characterizations that express a material’s resistance to deformation. In this paper, artificial intelligence (AI)-based models—especially M5P, support vector machine (SVM), and Gaussian process regression (GPR) algorithms—are built for R-value evaluation that meets the high precision and rapidity requirements in highway engineering. The dataset of this study comprises seven parameters: hydrated lime-activated rice husk ash, liquid limit, plastic limit, plasticity index, optimum moisture content, and maximum dry density. The available data are divided into three parts: training set (70%), test set (15%), and validation set (15%). The output (i.e., R-value) of the developed models is evaluated using the performance measures coefficient of determination (R2), mean absolute error (MAE), relative squared error (RSE), root mean square error (RMSE), relative root mean square error (RRMSE), performance indicator (ρ), and visual framework (Taylor diagram). GPR is concluded to be the best performing model (R2, MAE, RSE, RMSE, RRMSE, and ρ equal to 0.9996, 0.0258, 0.0032, 0.0012, 0.0012, and 0.0006, respectively, in the validation phase), very closely followed by SVM, and M5P. The application used for the aforementioned approaches for predicting the R-value is also compared with the recently developed artificial neural network model in the literature. The analysis of performance measures for the R-value dataset demonstrates that all the AI-based models achieved comparatively better and reliable results and thus should be encouraged in further research. Sensitivity analysis suggests that all the input parameters have a significant influence on the output, with maximum dry density being the highest.