High Point Density Research Articles

Study of the application of UAV and LiDAR to topographic survey of mountainous forest areas – Case study of Win 3 Wind Power Plant Project, Huong Hoa District, Quang Tri Province, Vietnam

Abstract Unmanned Aerial Vehicle (UAV) technology in conjunction with aerial photography and aerial Light Detection and Ranging (LiDAR) was used for topographic surveying of the Win 3 Wind Power Project area, a complex mountainous terrain situated in the Huong Hoa district of Quang Tri province, Vietnam with the primary objective to produce a map with a scale of 1:1,000 with a contour interval of 1.0 m, accompanied by 3D point clouds, a Digital Elevation Model (DEM), and an orthomosaic. Employing the UAV LiDAR system of DJI Matrice 300 RTK, we conduct ed survey flights, achieving a high point density laser scanning of 229 points/m2 and capturing ultra-high-resolution imagery at 4.22 cm/pixel. The integration of aerial LiDAR and photographic data results in a rich and detailed information repository regarding the surveyed terrain demonstrating that UAV LiDAR technology is a robust tool for geographic data collection, particularly in challenging terrain conditions in Vietnam. We also evaluated the accuracy of the created topographic map, achieving results that surpass the technical requirements and quality standards of the project. This underscores the potential and effectiveness of UAV LiDAR technology in the field of large-scale geographic data collection and mapping in Vietnam.

Clustering in point processes on linear networks using nearest neighbour volumes

This study introduces a novel method specifically designed to detect clusters of points within linear networks. This method extends a classification approach used for point processes in spatial contexts. Unlike traditional methods that operate on planar spaces, our approach adapts to the unique geometric challenges of linear networks, where classical properties of point processes are altered, and intuitive data visualisation becomes more complex. Our method utilises the distribution of the Kth nearest neighbour volumes, extending planar-based clustering techniques to identify regions of increased point density within a network. This approach is particularly effective for distinguishing overlapping Poisson processes within the same linear network. We demonstrate the practical utility of our method through applications to road traffic accident data from two Colombian cities, Bogota and Medellin. Our results reveal distinct clusters of high-density points in road segments where severe traffic accidents (resulting in injuries or fatalities) are most likely to occur, highlighting areas of increased risk. These clusters were primarily located on major arterial roads with high traffic volumes. In contrast, low-density points corresponded to areas with fewer accidents, likely due to lower traffic flow or other mitigating factors. Our findings provide valuable insights for urban planning and road safety management.

Challenges and Opportunities of Sentinel-1 InSAR for Transport Infrastructure Monitoring

Monitoring displacement at transport infrastructure using Sentinel‑1 Interferometric Synthetic Aperture Radar (InSAR) faces challenges due to the sensor’s medium spatial resolution, which limits the pixel coverage over the infrastructure. Therefore, carefully selecting coherent pixels is crucial to achieve a high density of reliable measurement points and to minimize noisy observations. This study evaluates the effectiveness of various pixel selection methods for displacement monitoring within transport infrastructures. We employ a two-step InSAR time series processing approach. First, high-quality first-order pixels are selected using temporal phase coherence (TPC) to estimate and correct atmospheric contributions. Then, a combination of different pixel selection methods is applied to identify coherent second-order pixels for displacement analysis. These methods include amplitude dispersion index (ADI), TPC, phase linking coherence (PLC), and top eigenvalue percentage (TEP), targeting both point-like scatterer (PS) and distributed scatterer (DS) pixels. Experiments are conducted in two case studies: one in Germany, characterized by dense vegetation, and one in Spain, with sparse vegetation. In Germany, the density of measurement points was approximately 30 points/km², with the longest segment of the infrastructure without any coherent pixels being 2.8 km. In Spain, the density of measurement points exceeded 500 points/km², with the longest section without coherent pixels being 700 meters. The results indicate that despite the challenges posed by medium-resolution data, the sensor is capable of providing adequate measurement points when suitable pixel selection methods are employed. However, careful consideration is necessary to exclude noisy pixels from the analysis. The findings highlight the importance of choosing a proper method tailored to infrastructure characteristics. Specifically, combining TPC and PLC methods offers a complementary set of pixels suitable for displacement measurements, whereas ADI and TEP are less effective in this context. This study demonstrates the potential of Sentinel‑1 InSAR for capturing both regional-scale and localized displacements at transport infrastructure.

Direct synthesis of polar functionalized linear low density polyethylene (LLDPE) using slow-chain-walking palladium catalysts

Linear low-density polyethylene (LLDPE) is the most widely produced type of polyethylene resin. In this work, we detailed the synthesis and analysis of two bulky α-diimine Pd(II) complexes. These Pd(II) complexes can efficiently synthesize LLDPE-type polyethylene materials (ρ = 0.91–0.93 g/cm3) through ethylene slow-chain-walking polymerization. More interestingly, benefiting from the cationic bulky Pd(II) complexes, polar-functionalized LLDPEs (ρ = 0.91–0.92 g/cm3) are also produced by copolymerization of ethylene with various polar monomers such as 10-methyl undecanoate, methyl acrylate, 10-undecenoic acid. The copolymerization maintains the slow-chain-walking polymerization feature, resulting in a polar copolymer with a high melting point and low branching density. The results of 13C NMR spectral also reveal that polar groups are mainly located at the end of the branches in polar-functionalized LLDPEs.

VOXEL Video Streaming Over Wireless Networks

High-density point clouds expressing attractive 3D images are attracting attention. These gigantic media require large bandwidth allocations, making them problematic to stream to resource-constrained hand-held devices. This paper proposes a method for point cloud compression and streaming of large point clouds using a web server. Storing large point cloud videos on a web server allows users to publish data sets without using additional applications or sending large amounts of data ahead of time. HTTP/2 improves transfer efficiency by compressing headers into binary format and reduces latency by sending many multiplexed requests using a single TCP session. The large size of the point cloud data and the current limitations of wireless channels demand point cloud compression methods. MATLAB is used for the purpose of analysis, at results show an average compression ratio of 30:1. Our proposed system achieves lower average latency between successive frames resulting in an increased frame rate to 72.46 fps when streaming 4M points, and 82.51 fps in the case of streaming 800K points compared to much lower rates in conventional work.

Mapping temperate old-growth forests in Central Europe using ALS and Sentinel-2A multispectral data

Old-growth forests are essential to preserve biodiversity and play an important role in sequestering carbon and mitigating climate change. However, their existence across Europe is vulnerable due to the scarcity of their distribution, logging, and environmental threats. Therefore, providing the current status of old-growth forests across Europe is essential to aiding informed conservation efforts and sustainable forest management. Remote sensing techniques have proven effective for mapping and monitoring forests over large areas. However, relying solely on remote sensing spectral or structural information cannot capture comprehensive horizontal and vertical structure complexity profiles associated with old-growth forest characteristics. To overcome this issue, we combined spectral information from Sentinel-2A multispectral imagery with 3D structural information from high-density point clouds of airborne laser scanning (ALS) imagery to map old-growth forests over an extended area. Four features from the ALS data and fifteen from Sentinel-2A comprising raw band (spectral reflectance), vegetation indices (VIs), and texture were selected to create three datasets used in the classification process using the random forest algorithm. The results demonstrated that combining ALS and Sentinel-2A features improved the classification performance and yielded the highest accuracy for old-growth class, with an F1-score of 92% and producer’s and user’s accuracies of 93% and 90%, respectively. The findings suggest that features from ALS and Sentinel-2A data sensitive to forest structure are essential for identifying old-growth forests. Integrating open-access satellite imageries, such as Sentinel-2A and ALS data, can benefit forest managers, stakeholders, and conservationists in monitoring old-growth forest preservation across a broader spatial extent.

Hydrogels with high sacrifice efficiency of sacrificial bonds and with high strength and toughness due to dense entanglements of polymer chains

Introducing sacrificial bonds is a common method for increasing the toughness of hydrogels. Many sacrificial bonds have been extensively investigated, but the sacrifice efficiency has never been studied. In this study, polyacrylamide hydrogels with highly entangled polymer chains containing carboxyl–zirconium (–COO−–Zr4+) sacrificial bonds are prepared to study the effect of polymer chain entanglement on the sacrificial bond efficiency. Unlike chemical crosslinking points, the dense physical entanglements do not affect the toughness (∼43 MJ/m3) of hydrogels but significantly improve the tensile strength (by two times) and Young’s modulus (by six times). Physical entanglements enable the chains to slide and adjust the network structure under stress, which enables more polymer chains and sacrificial bonds to participate in the deformation process. Therefore, dense entanglements will greatly improve the sacrifice efficiency. However, a high density of chemical crosslinking points will limit the improvement in the sacrifice efficiency, which is attributed to the sliding limitations because of physical entanglement. The highly entangled polyacrylamide hydrogels toughened by –COO−–Zr4+ have an excellent load-bearing capacity. This study provides a novel strategy for designing hydrogels with ultra-high strength and toughness, which paves the way for the development of many hydrogels used in engineering materials.

Monocular Metasurface for Structured Light Generation and 3D Imaging with a Large Field-of-View.

Structured light three-dimensional (3D) imaging technology captures the geometric information on 3D objects by recording waves reflected from the objects' surface. The projection angle and point number of the laser dots directly determine the field-of-view (FOV) and the resolution of the reconstructed image. Conventionally, diffractive optical elements with micrometer-scale pixel size have been used to generate laser dot arrays, leading to limited FOV and point number within the projection optical path. Here, we theoretically put forward and experimentally demonstrate a monocular geometric phase metasurface composed of deep subwavelength meta-atoms to generate a 10 798 dot array within an FOV of 163°. Attributed to the vast number and high-density point cloud generated by the metasurface, the 3D reconstructed results showcase a maximum relative error in depth of 5.3 mm and a reconstruction error of 6.07%. Additionally, we propose a spin-multiplexed metasurface design method capable of doubling the number of lattice points. We demonstrate its application in the field of 3D imaging through experiments, where the 3D reconstructed results show a maximum relative depth error of 0.44 cm and a reconstruction error of 2.78%. Our proposed metasurface featuring advanced point cloud generation holds substantial potential for various applications such as facial recognition, autonomous driving, virtual reality, and beyond.

H+ implantation induced defects distribution in 4H-SiC single crystal film fabricated by Crystal-Ion-Slicing and its effects on electrical behavior: A multiple characterization study

The composite substrate of SiC single crystal films fabricated by Crystal-Ion-Slicing (CIS) is acknowledged as an excellent material platform for the realization of electronic devices with specific functions. In CIS technology, the transferred monocrystalline films can retain good monocrystalline quality close to the virgin monocrystalline quality, but the lattice damage caused by ion implantation and crystal cleavage is still not negligible. The distribution of H+ implantation induced defects is studied in this paper, according to which the transferred film can be divided into lattice distortion region, moderately damaged region and application region. The lattice distortion region has serious lattice defects and stacking faults, while the moderately damaged layer is dominated by high lattice stress and high density point defects. For application region, in addition to the defect morphology, its influence on the electrical behavior is also investigated, which indicate the defects are mainly vacancy defects introduced by the H+ penetration, which act as carrier traps and recombination centers in electronic behavior. The research work is helpful to further understand the microstructure and defect distribution of the films of various materials fabricated by the common technology CIS, and can provide a theoretical basis for further defect layer removal and defect repair.

Fusion of Multiple Basic Element Features for Airborne LiDAR in-house Surveys

When using airborne LiDAR point clouds for city modelling and road extraction, point cloud classification is a crucial step. There are numerous ways for classifying point clouds, but there are still issues like redundant multi-dimensional feature vector data and poor point cloud classification in intricate situations. A point cloud classification method built on the fusing of multikernel feature vectors is suggested as a solution to these issues. The technique employs random forest to classify point cloud data by merging colour information, and it extracts feature vectors based on point primitives and object primitives, respectively. In this study, a densely populated area was chosen as the study area. Light airborne LIDAR mounted on a delta wing was used to collect point cloud data at a low altitude (170 m) over a dense cross-course. The point cloud data were then combined, corrected, and enhanced with texture data, and the houses were vectorized on the point cloud. The accuracy of the results was then assessed. With a median inaccuracy of 4.8 cm and a point cloud data collection rate of 83.3%, using airborne LIDAR to measure house corners can significantly lighten the labour associated with external house corner measurements.This test extracts the texture information of point cloud data through the efficient processing of high-density point cloud data, providing a reference for the application of texture information of airborne LIDAR data and a clear understanding of its accuracy.

W-doped concentration improved the hardness, elastic modulus, elastic anisotropy and thermodynamic properties of ZrC ultrahigh temperature ceramics

Although ZrC is a promising ultrahigh temperature ceramics due to the high melting point, high strength, high hardness and low density, the improvement of mechanical properties and thermal stability is very significance for it's ultrahigh temperature application. To improve the mechanical and thermodynamic properties of ZrC ultrahigh temperature ceramics, we apply the first-principles calculations to study the influence of W-doped concentration on the structural stability, hardness, elastic modulus, elastic anisotropy and thermodynamic properties of the cubic ZrC ceramics. The calculated results show that three W-doped ZrC are thermodynamic stability. In particular, the calculated impurity formation energy of the W-doped ZrC increases with increasing W concentration. Importantly, the calculated elastic modulus of the W-doped ZrC increases with increasing W concentration. Although the hardness of W-doped ZrC is bigger than the parent ZrC, the hardness of W-3.125 at.% is higher than the W-doped ZrC with the high concentration of W. Naturally, the high hardness and elastic modulus of the W-doped ZrC is that the metal W enhances the localized hybridization between Zr and C, which is demonstrated by the change of ZrC bond. In addition, the parent ZrC and the W-doped ZrC show excellent elastic anisotropy. Furthermore, the calculated Debye temperature of W-3.125 at.% ZrC is higher than the parent ZrC and the other W-doped ZrC. Therefore, we predict that the W-3.125 at.% ZrC shows excellent mechanical properties and better thermodynamic properties in comparison to the other W-doped ZrC ceramics.

A Hybrid Method for Individual Tree Detection in Broadleaf Forests Based on UAV-LiDAR Data and Multistage 3D Structure Analysis

Individual tree detection and segmentation in broadleaf forests have always been great challenges due to the overlapping crowns, irregular crown shapes, and multiple peaks in large crowns. Unmanned aerial vehicle (UAV)-borne light detection and ranging (LiDAR) is a powerful tool for acquiring high-density point clouds that can be used for both trunk detection and crown segmentation. A hybrid method that combines trunk detection and crown segmentation is proposed to detect individual trees in broadleaf forests based on UAV-LiDAR data. A trunk point distribution indicator-based approach is first applied to detect potential trunk positions. The treetops extracted from a canopy height model (CHM) and the crown segments obtained by applying a marker-controlled watershed segmentation to the CHM are used to identify potentially false trunk positions. Finally, the three-dimensional structures of trunks and branches are analyzed at each potentially false trunk position to distinguish between true and false trunk positions. The method was evaluated on three plots in subtropical urban broadleaf forests with varying proportions of evergreen trees. The F-score in three plots ranged from 0.723 to 0.829, which are higher values than the F-scores derived by a treetop detection method (0.518–0.588) and a point cloud-based individual tree segmentation method (0.479–0.514). The influences of the CHM resolution (0.25 and 0.1 m) and the data acquisition season (leaf-off and leaf-on) on the final individual tree detection result were also evaluated. The results indicated that using the CHM with a 0.25 m resolution resulted in under-segmentation of crowns and higher F-scores. The data acquisition season had a small influence on the individual tree detection result when using the hybrid method. The proposed hybrid method needs to specify parameters based on prior knowledge of the forest. In addition, the hybrid method was evaluated in small-scale urban broadleaf forests. Further research should evaluate the hybrid method in natural forests over large areas, which differ in forest structures compared to urban forests.

Towards monitoring of critical illness via the detection of histones with extended gate field-effect transistor sensors

Extracellular histone proteins in the blood indicate a heightened risk of morbidity after trauma or in major illnesses such as sepsis. We present the development of an aptasensor for histone detection with an extended gate field-effect transistor (EGFET) configuration, which benefits from low power consumption, rapid response, and compatibility with miniaturized gold electrodes. Histones have a high isoelectric point and charge density, which cause them to physically adsorb to non-specific elements of the sensor that have available electrostatic charges. To combat this, the sensing surface is formed with a thiol-modified, high-affinity and histone-specific RNA aptamer sequence and by co-immobilizing with poly(ethylene glycol) methyl ether thiol (PEG) as a blocking agent. Surface plasmon resonance (SPR) is used to analyze aptamer and PEG immobilization strategies, confirm histone binding, and calculate kinetic binding constants. Through comparison of different blocking agents and time-dependent preparation, the ideal equilibrium dissociation constant (KD) is estimated to be below 200 pM, which is the upper range of extracellular histone concentrations in critically ill patients with high mortality. The EGFET sensitivity of the optimized aptasensor is 6.65 mV/decade concentration change for histone H4 with a physiologically relevant 5 pM limit of detection. Selectivity tests with 100 nM bovine serum albumin (BSA) demonstrate a signal response that is 13-fold smaller than for histones. This EGFET aptasensor platform is suitable for future point-of-care monitoring of histone levels in critically ill patients, thus permitting the early detection of increased risk and the need for more aggressive interventional measures to prevent mortality.

RMS-FlowNet++: Efficient and Robust Multi-scale Scene Flow Estimation for Large-Scale Point Clouds

The proposed RMS-FlowNet++ is a novel end-to-end learning-based architecture for accurate and efficient scene flow estimation that can operate on high-density point clouds. For hierarchical scene flow estimation, existing methods rely on expensive Farthest-Point-Sampling (FPS) to sample the scenes, must find large correspondence sets across the consecutive frames and/or must search for correspondences at a full input resolution. While this can improve the accuracy, it reduces the overall efficiency of these methods and limits their ability to handle large numbers of points due to memory requirements. In contrast to these methods, our architecture is based on an efficient design for hierarchical prediction of multi-scale scene flow. To this end, we develop a special flow embedding block that has two advantages over the current methods: First, a smaller correspondence set is used, and second, the use of Random-Sampling (RS) is possible. In addition, our architecture does not need to search for correspondences at a full input resolution. Exhibiting high accuracy, our RMS-FlowNet++ provides a faster prediction than state-of-the-art methods, avoids high memory requirements and enables efficient scene flow on dense point clouds of more than 250K points at once. Our comprehensive experiments verify the accuracy of RMS-FlowNet++ on the established FlyingThings3D data set with different point cloud densities and validate our design choices. Furthermore, we demonstrate that our model has a competitive ability to generalize to the real-world scenes of the KITTI data set without fine-tuning.

Improving LiDAR height precision in urban environment: Low-cost GNSS ranging prototype for post-mission airborne laser scanning enhancement

Although Light Detection and Ranging (LiDAR) technology is currently one of the most efficient methods for acquiring high-density point cloud, there are still challenges in terms of data reliability. In particular, the accuracy assessment of LiDAR data, especially in the height component, is one of the main issues in this context. This study introduces a rapid and cost-effective platform to improve the accuracy and precision of LiDAR data by integrating high-density GNSS-Ranging measurements with LiDAR data. The platform offers the capability to rapidly collect a significant number of network real time kinematic (NRTK) points with centimetric precision. A continuous correction surface is proposed to integrate the platform and LiDAR data, resulting in improved accuracy for all ground-class LiDAR data. Evaluation using GNSS benchmarks and NRTK checkpoints showed a significant reduction in LiDAR height errors after applying the correction surface. The root mean squares error (RMSE) decreased from 18.5 cm to 8.2 cm when compared to GNSS benchmarks and from 17.4 cm to 5.3 cm for approximately 1000 NRTK control points. The results indicate that collecting a large number of high-density GNSS ground targets and applying a correction surface to LiDAR height data significantly enhance the accuracy and precision of the LiDAR extracted products.

Building Extraction from LiDAR Point Clouds Based on Revised RandLA-Net

Abstract. 3D building models is crucial for applications in smart cities. Automatic reconstruction of 3D buildings has been investigated based on various data sources. Point clouds from airborne LiDAR scanners can be used to extract buildings data due to its high accuracy and point density. In this paper, we present a methodology to segment buildings and corresponding rooftop structure from point clouds. First, RandLA-Net, which is an efficient and lightweight neural network for semantic segmentation of large-scale point clouds, is revised and adopted for building segmentation. By implementing local feature aggregation of each point, RandLA-Net can effectively preserve geometric details in point clouds. Besides 3D coordinates of point clouds, we incorporated point attributes including pulse intensity and return numbers into the network as additional features. Feature normalizations are applied to the input features. To achieve a better result of the local feature aggregation, hyperparameters of the network are fine-tuned according to the density of points and building size. Based on the classified building point clouds, DBSCAN clustering algorithm is implemented for segmenting individual buildings. Elevation histogram analysis is conducted to determine optimal threshold values for delineating candidate rooftop point clouds of individual buildings. For the buildings with multiple rooftops, multiple elevation threshold values are necessary to extract corresponding rooftops or walls. Then DBSCAN is employed again for segmentation of individual rooftops and denoising of point clouds of each building. Finally, Alpha-shape analysis is applied based on adaptive threshold values to build the envelope of each rooftop. Experiments show that our implementation of building segmentation using RandLA-net achieves higher mean IoU (Intersection over Union) and better classification performance in building segmentation. ISPRS benchmark data was used in our experiment and our methodology produce results with accuracy of 90.79%.

3D scanning and modeling of highly detailed and geometrically complex historical architectural objects: the example of the Juma Mosque in Khiva (Uzbekistan)

The city centre of Khiva (Uzbekistan), called Itchan Kala, is an architectural complex included in the UNESCO list of tangible cultural heritage. One of the historic buildings in it is the Juma Mosque. It has a simple rectangular structure, but is very large and has 213 deeply carved wooden columns supporting the roof. The article presents the process, problems, and their solutions resulting from the implementation of 3D laser scanning of such highly detailed and geometrically complex historical architectural objects in the conditions of normal tourist traffic. The optimisation of scanning positions, scanning in situ implementation, as well as the processing of the acquired data and the construction of a 3D mesh model of the mosque interior are presented. It is pointed out that scanning such objects with high accuracy and density of measurement points causes major technical problems related to the workload, and the huge volume of data acquired and processed. The possibilities of making the 3D model available in digital space for the purpose of researching the appearance and geometry of the mosque, its individual columns, as well as popularising the monument are also discussed. It is highly probable that the scanning of the Juma Mosque's interior presented here was carried out for the first time in history.

A study of annual tree-wise LiDAR intensity patterns of boreal species observed using a hyper-temporal laser scanning time series

This study introduces the annual tree-wise intensity patterns of three boreal tree species, silver birch (Betula pendula Roth.), Scots pine (Pinus sylvestris L.), and Norway spruce (P,icea abies H. Karst.), observed from a long-term hyper-temporal point cloud dataset collected with a permanent laser scanning (LiDAR) station. An annual LiDAR intensity pattern refers to the trend of variations of tree-wise calibrated LiDAR intensity values over the course of a year, primarily linked to species-specific phenological characteristics. Such pattern was discovered from hyper-temporal (76 observations between April 2020 and April 2021) high resolution (0.01 m 3D point spacing at 100 m distance) point cloud observations acquired using a single wavelength (1550 nm) static LiDAR system installed on a climate monitoring tower at the Hyytiälä forest research station in central Finland. A set of experiments was designed to explore the interactions among tree-wise calibrated LiDAR intensity, species, data quality, and observation dates. Thus, to provide practical suggestions for expectable accuracies of tree species classification using diverse types of LiDAR systems and data acquisition strategies. It was revealed that by combining high spatial resolution (up to 100,000 points/m2) with bi-weekly observations (two scans per week except winter period), the average tree species classification accuracy reaches 96.8% for the three boreal tree species in the study area, where the few misclassified cases are attributed to inaccurate individual tree segmentation. This suggests that the annual tree-wise LiDAR intensity patterns, which reflects the tree-wise phenological dynamics, are robust and species-specific. When the spatial resolution was reduced to 5 points/ m2, the average accuracy of species classification among the three studied species reached 85% when using bi-temporal data from spring and autumn. Our analysis further indicates that arranging data acquisition with a two-week variability in late autumn and late spring could yield higher accuracy in species classification. Furthermore, our study revealed the importance of subcanopy information for species classification, specifically in distinguishing different coniferous species. When limited to upper canopy data (as in simulated airborne LiDAR scenarios), achieving a tree species classification accuracy of over 80% required either high point density (e.g. 8000 points/m2) combined with bi-temporal collection in a year, or low point density (5 points/ m2) with a dense time series (e.g., 76 time points within a year), which showcased the possibility to compensate low spatial resolution with high temporal resolution, and vice versa, in capturing species-specific phenological characteristics of trees.

Automated forest inventory: Analysis of high-density airborne LiDAR point clouds with 3D deep learning

Detailed forest inventories are critical for sustainable and flexible management of forest resources, to conserve various ecosystem services. Modern airborne laser scanners deliver high-density point clouds with great potential for fine-scale forest inventory and analysis, but automatically partitioning those point clouds into meaningful entities like individual trees or tree components remains a challenge. The present study aims to fill this gap and introduces a deep learning framework, termed ForAINet, that is able to perform such a segmentation across diverse forest types and geographic regions. From the segmented data, we then derive relevant biophysical parameters of individual trees as well as stands. The system has been tested on FOR-Instance, a dataset of point clouds that have been acquired in five different countries using surveying drones. The segmentation back-end achieves over 85% F-score for individual trees, respectively over 73% mean IoU across five semantic categories: ground, low vegetation, stems, live branches and dead branches. Building on the segmentation results our pipeline then densely calculates biophysical features of each individual tree (height, crown diameter, crown volume, DBH, and location) and properties per stand (digital terrain model and stand density). Especially crown-related features are in most cases retrieved with high accuracy, whereas the estimates for DBH and location are less reliable, due to the airborne scanning setup.

A neighborhood-based robust clustering algorithm using Apollonius function kernel

Density Peak Clustering (DPC) is a useful algorithm for grouping data points based on their closeness and density. It is widely used in various areas of computer science. It can detect the peak density in each period of the clustering steps and assign the rest of the data to the next steps. Peak density clustering analysis is able to discover clusters of arbitrary shapes, but usually, when the data set is large, it is computationally expensive and difficult to apply to high-dimensional data. When the distance between clusters is significantly different, it is unable to distinguish adjacent clusters with different densities. To deal with such issues, a new clustering algorithm based on neighborhoods and the Apollonius kernel is proposed. The proposed algorithm can be implemented with four steps. 1) Estimation of natural neighbor, 2) Identification of core points, 3) Merging the cores of a cluster and separating the cores of different clusters, 4) Using the Apollonius kernel between the core points and the farthest points to assign the remaining points to the clusters and find the noise points. The local density estimation process based on the natural neighborhood of high-density points and their integration, as well as border point classification, allows the proposed method to reveal the local cluster structure. As a result, the adjoining clusters with distinctive densities are recognized, and the computation is substantially reduced. The proposed method is validated on some synthetic data sets and the UCI data set experimentally and has a high 10% accuracy.