High-resolution Light Detection And Ranging Research Articles

Three-Dimensional Mapping of Habitats Using Remote-Sensing Data and Machine-Learning Algorithms

Progress toward habitat protection goals can effectively be performed using satellite imagery and machine-learning (ML) models at various spatial and temporal scales. In this regard, habitat types and landscape structures can be discriminated against using remote-sensing (RS) datasets. However, most existing research in three-dimensional (3D) habitat mapping primarily relies on same/cross-sensor features like features derived from multibeam Light Detection And Ranging (LiDAR), hydrographic LiDAR, and aerial images, often overlooking the potential benefits of considering multi-sensor data integration. To address this gap, this study introduced a novel approach to creating 3D habitat maps by using high-resolution multispectral images and a LiDAR-derived Digital Surface Model (DSM) coupled with an object-based Random Forest (RF) algorithm. LiDAR-derived products were also used to improve the accuracy of the habitat classification, especially for the habitat classes with similar spectral characteristics but different heights. Two study areas in the United Kingdom (UK) were chosen to explore the accuracy of the developed models. The overall accuracies for the two mentioned study areas were high (91% and 82%), which is indicative of the high potential of the developed RS method for 3D habitat mapping. Overall, it was observed that a combination of high-resolution multispectral imagery and LiDAR data could help the separation of different habitat types and provide reliable 3D information.

Inside Late Bronze Age Settlements in NE Romania: GIS-Based Surface Characterization of Ashmound Structures Using Airborne Laser Scanning and Aerial Photography Techniques

The identification and delineation, through aerial photography, of the archaeological structures that present temporal resolution, as well as their characterization based on high-resolution LiDAR (Light Detection and Ranging)-derived DEMs (Digital Elevation Models) are modern techniques widely used in the archaeological prospecting of various landscapes. In this study, we present an application of Airborne Laser Scanning (ALS) and aerial photography (AP) techniques, used in order to compute geomorphometric indices specific to the ashmound structures of Late Bronze Age (LBA) archaeological sites that are visible on the soil surface. The necessity of determining the ashmounds’ geoarchaeological description stems from the fact that despite the majority of archaeologists weighing in on the subject, there is still no accepted explanation regarding their initial functionality. Thus, we believe that the GIS-based high-resolution characterization of 200 ashmound features identified in 21 Noua Culture (NC) archaeological sites will contribute to a better understanding of the ashmounds’ functionality and evolution in the heterogeneous landscape of the study area (NE Romania). Therefore, various shape indices, such as the area (A), perimeter (P), length (L), form factor (RF), circularity ratio (RC), and elongation ratio (RE) were computed for microlevel characterizations of the visible ashmounds’ structures. Additionally, LiDAR-derived DEMs with a 0.5 m resolution were used to generate more surface characteristics such as the slope (S) and hypsometric indices (HI). The outcomes indicate that the ashmounds have relatively diverse shapes (an RF range from 0.37 to 0.77; a RC range from 0.79 to 0.99; a RE range from 0.68 to 0.99), and the micro-relief slightly varies from positive to negative landforms (HI range from 0.34 to 0.61) depending on the erosion intensity (S range from 1.17° to 19.69°) and anthropogenic impact (e.g., current land use and agriculture type). Furthermore, each morphometric parameter is an indicator for surface processes, aiding in the identification of the geomorphologic and surface-erosion aspects that affect the archaeological remains, contributing to the assessment of the conservation status of the ashmound structures within the current landscape configuration. In this regard, this article presents and discusses the remote sensing (RS) techniques used, as well as the morphometric data obtained, exploring the implications of our findings for a better characterization of the NC in Romania.

Defining multi-scale surface roughness of a coral reef using a high-resolution LiDAR digital elevation model

A decline in coral reef surface roughness may indicate that a coral reef ecosystem is approaching functional collapse. This is because surface roughness underpins many critical ecosystem metrics, such as, live coral cover and high fish biomass. Yet, we understand very little about how surface roughness changes at ecosystem scales due to the limited spatial extents of coral reef ecological surveys. Here, we investigate the surface roughness of a coral reef across multiple spatial scales on a high-resolution Light Detection and Ranging (LiDAR) derived digital elevation model (DEM, 0.25 m cells). We found that complex features dominated regions with high roughness (e.g., spur and groove systems on the reef slope) and increased the average roughness of geomorphic zones that are otherwise flat and featureless (e.g., lagoonal patch reefs in the sandy deep lagoon). Each geomorphic zone had a unique signature with high values of roughness usually expressed over finer spatial scales and, conversely, lower values of roughness observed over broader spatial scales. We allocated each geomorphic zone into one of four roughness equivalent habitats (REHs 1–4) that share similar surface roughness properties but are not necessarily geographically contiguous regions. These results identify fine scale features and geomorphic zones that are important for providing roughness to coral reef systems, such as spur and groove systems and the fore-reef slope respectively. They also suggest that morphodynamic processes operating at broader spatial scales influence the physical structure of coral reef ecosystems. The quantification of coral reef surface roughness is becoming increasingly important due to the wider availability of high (<1 m) and hyper (<0.1 m) resolution DEMs. Our approach and findings here can be used to provide greater surface textural information for common coral reef geomorphic zones and aid future management and research efforts, such as, monitoring of coral reef ecosystem response to environmental change and reef restoration and adaptation programs.

Development of the high angular resolution 360° LiDAR based on scanning MEMS mirror

Light detection and ranging (LiDAR) using various operational principles has been applied in many fields, e.g., robotics navigation, autonomous vehicles, unmanned aerial flyers, land surveying, etc. The multichannel LiDAR system is of great importance in the field of autonomous driving due to its larger field of view (FoV). However, the number of transceivers limits the vertical angular resolution of multichannel LiDAR systems and makes them costly. On the other hand, the emergence of microelectromechanical systems (MEMS) mirrors may provide a highly promising solution to a low-cost, high angular resolution LiDAR system. We have demonstrated a MEMS mirror-based 360° LiDAR system with high angular resolution and will present the detailed design process and obtained experimental results in this paper. With the combination of the MEMS mirror and a rotation platform for the LiDAR system, a 360° × 8.6° (horizontal × vertical) FoV was achieved. Compared with existing commercial multichannel 360° LiDAR systems, our system has 13.8 times better angular resolution than the Velodyne HDL-64 LiDAR sensor. The experimental results verified an excellent performance of 0.07° × 0.027° (horizontal × vertical) angular resolution, which enhances the panoramic scanning and imaging capability of the LiDAR system, potentially providing more accurate 3D scanning applications in areas such as autonomous vehicles, indoor surveying, indoor robotics navigation, etc.

Winter adverse driving dataset for autonomy in inclement winter weather

The availability of public datasets with annotated light detection and ranging (LiDAR) point clouds has advanced autonomous driving tasks, such as semantic and panoptic segmentation. However, there is a lack of datasets focused on inclement weather. Snow and rain degrade visibility and introduce noise in LiDAR point clouds. In this article, summarize a 3-year winter weather data collection effort and introduce the winter adverse driving dataset. It is the first multimodal dataset featuring moderate to severe winter weather—weather that would cause an experienced driver to alter their driving behavior. Our dataset features exclusively events with heavy snowfall and occasional white-out conditions. Data are collected using high-resolution LiDAR, visible as well as near infrared (IR) cameras, a long wave IR camera, forward-facing radio detection and ranging, and Global Navigation Satellite Systems/Inertial Measurement Unit units. Our dataset is unique in the range of sensors and the severity of the conditions observed. It is also one of the only data sets to focus on rural and semi-rural environments. Over 36 TB of adverse winter data have been collected over 3 years. We also provide dense point-wise labels to sequential LiDAR scans collected in severe winter weather. We have labeled and will make available around 1000 sequential LiDAR scenes, amounting to over 7 GB or 3.6 billion labeled points. This is the first point-wise semantically labeled dataset to include falling snow.

Enabling 3-D Object Detection With a Low-Resolution LiDAR

Light detection and ranging (LiDAR) has been widely used in autonomous vehicles for perception and localization. However, the cost of a high-resolution LiDAR is still prohibitively expensive, while its low-resolution counterpart is much more affordable. Therefore, using low-resolution LiDAR for autonomous driving is an economically viable solution, but the point cloud sparsity makes it extremely challenging. In this letter, we propose a two-stage neural network framework that enables 3-D object detection using a low-resolution LiDAR. Taking input from a low-resolution LiDAR point cloud and a monocular camera image, a depth completion network is employed to produce dense point cloud that is subsequently processed by a voxel-based network for 3-D object detection. Evaluated with KITTI dataset for 3-D object detection in bird-eye view (BEV), the experimental result shows that the proposed approach performs significantly better than directly applying the 16-line LiDAR point cloud for object detection. For both easy and moderate cases, our 3-D vehicle detection results are close to those using 64-line high-resolution LiDARs.

Energy Efficient Approximate 3D Image Reconstruction

We demonstrate an efficient and accelerated parallel, sparse depth reconstruction framework using compressed sensing (compressed sensing (CS)) and approximate computing. Employing data parallelism for rapid image formation, the depth image is reconstructed from sparsely sampled scenes using convex optimization. Coupled with faster imaging, this sparse sampling reduces significantly the projected laser power in active systems such as light detection and ranging (LiDAR) to allow eye safe operation at longer range. We also demonstrate how reduced precision is leveraged to reduce the number of logic units in field-programmable gate array (FPGA) implementations for such sparse imaging systems. It enables significant reduction in logic units, memory requirements and power consumption by over 80% with minimal impact on the quality of reconstruction. To further accelerate processing, pre-computed, important components of the lower-upper (LU) decomposition and other linear algebraic computations are used to solve the convex optimization problems. Our methodology is demonstrated by the application of the alternating direction method of multipliers (ADMM) and proximal gradient descent (PGD) algorithms. For comparison, a fully discrete least square reconstruction method ( <inline-formula><tex-math notation="LaTeX">$d$</tex-math></inline-formula> Sparse) is also presented. This demonstrates the feasibility of novel, high resolution, low power and high frame rate LiDAR depth imagers based on sparse illumination for use in applications where resources are strictly limited.

Deglaciation history and subsequent lake dynamics in the Siljan region, south‐central Sweden, based on new LiDAR evidence and sediment records

AbstractThe Siljan region hosts Europe's largest impact structure. The high‐relief landscape, with a central granite dome bordered by lake basins, contains an array of glacial and shore‐level landforms. We investigated its deglaciation history by mapping and analysing landforms on high resolution LiDAR (light detection and ranging)‐based digital surface models coupled with well‐dated sediment successions from peat and lake sediment cores. The granite dome and bordering areas are characterized by streamlined terrain and ribbed moraine with a streamlined overprint. These suggest an ice‐flow direction from north‐northwest (NNW) with wet‐based thermal conditions prior to deglaciation. During its retreat, the ice sheet was split into thinner plateau ice and thicker basin ice. Sets of low‐gradient glaciofluvial erosion channels suggest intense ice‐lateral meltwater drainage across gradually ice‐freed slopes, while ‘down‐the‐slope’ erosion channels and eskers show meltwater drainage from stagnated plateau ice. Thick basin ice receded with a subaqueous margin across the deep Siljan–Orsasjön Basin c. 10,700–10,500 cal. bp. During ice recession the ingression of the Baltic Ancylus Lake led to diachronous formation of highest shoreline marks, from ~207 m in the south to ~220 m above sea level (a.s.l.) in the north. Differential uplift resulted in shallowing of the water body, which led to the isolation of the Siljan–Orsasjön Basin from the Baltic Basin at c. 9800 cal. bp. The post‐isolation water body – the ‘Ancient Lake Siljan’ – was drained through the ancient Åkerö Channel with a water level at 168–169 m a.s.l. during c. 1000 years. A later rerouting of the outlet to the present course was initiated at c. 8800 cal. bp, which led to a lake‐level lowering of 6–7 m to today's level of Lake Siljan (~162 m a.s.l.). This study shows the strength of an integrated methodological approach for deciphering the evolution of a complex landscape, combining highly resolved geomorphological analysis with well‐dated sediment successions.

High-resolution LiDAR using random modulated continuous wave with a low code rate

Light detection and ranging (LiDAR) is essential to improving the reliability of autopilot technology. In this paper, we propose a new laser-ranging method for three-dimensional micro-electro-mechanical system LiDAR. First, a mathematical model is presented and the analytical solution for real-time calculation derived. Then, the mixed M-sequence and data code modulates the laser diode by pulse amplitude modulation. Finally, the resolution is enhanced by the matched filter and delay-locked loop technique, lowering the requirement of the code rate and analog-to-digital-converter sampling rate. Analytical and experimental results demonstrate that a distance precision of less than 2 cm and acquisition time within 8 ns (125 MHz) can be achieved, while the processing gain reaches 48 dB. In addition, this enhanced performance is maintained over long distances without periodic ambiguities, making it ideal for range imaging.

Review on the Application of Airborne LiDAR in Active Tectonics of China: Dushanzi Reverse Fault in the Northern Tian Shan

High-resolution topographic data are fundamental for active tectonics studies. Within the last 2 decades, the airborne light detection and ranging (LiDAR) system provided a solution for the accurate and efficient acquisition of detailed geomorphic features. The use of LiDAR data for the identification of microstructural and geomorphic features, fault zone activity analysis, and earthquake disaster assessment remains challenging and has been the focus of active tectonics studies. Based on the LiDAR data of Dushanzi anticline–a reverse fault zone in Xinjiang, China, our group carried out a significant number of active tectonic research studies. By reviewing the specific content of these works, we summarized the main application of LiDAR for a specific structure, the Dushanzi Reverse Fault in the northern Tianshan. In addition, other applications of LiDAR in active tectonics are summarized in this paper. These studies show that high-resolution LiDAR data facilitate detailed studies of the fault activity and paleoseismicity. We hope more researchers can realize the advantages of LiDAR technology in active tectonics research and apply LiDAR technology into their own practical work, so as to promote the development of active tectonics research.

Potential and limitations of LiDAR altimetry in archaeological survey. Copper Age and Bronze Age settlements in southern Iberia

Potential and limitations of LiDAR altimetry in archaeological survey. Copper Age and Bronze Age settlements in southern Iberia

Low-Timing-Jitter GHz-Gated InGaAs/InP Single-Photon Avalanche Photodiode for LIDAR

InGaAs/InP single-photon avalanche diodes (SPADs) attract increasing attention for the infrared light detection and ranging (LIDAR) due to the low-cost and compact-size construction. Although constant endeavor has been dedicated to improve the dead time and timing jitter, it remains challenging to implement high-speed and high-resolution LIDAR. Here, we develop a high-performance GHz-gated InGaAs/InP APD operated in the quasi-free-running mode, while its gating signal is unlocked from the light source. The used low-pass filtering technique enables us to continuously adjust the repetition frequency of the gate for carefully optimizing photon detection efficiency (PDE) and timing jitter. Increasing the gates’ repetition frequency from 1 GHz to 2.75 GHz, the PDE improves dramatically from 2.02% to 3.30%, while the timing jitter drops from 156 ps to 114 ps. Furthermore, we verify that changes in the PDE have little effect on the timing characteristics, making the SPAD more suitable for LIDAR with large dynamic range. Meanwhile, the dead time is about 4 ns, which is much shorter than that of the SPAD working in the free-running mode, guaranteeing rapid mapping of targets. Finally, as a proof-of-principle demonstration, the InGaAs/InP SPAD is used in a time-of-flight LIDAR system, which demonstrates a sub-centimeter depth resolution directly.

Landslide inventory and characteristics, based on LiDAR scanning and optimised field investigations in the Kutina area, Croatia

This paper presents the preliminary results of analyses of landsliding processes derived from detailed LiDAR (Light Detection and Ranging) scans supported by field prospection on the south-western slopes of Mt. Moslavačka gora, in the wider Kutina area. This area is known for frequent landslides, but dedicated regional landslide research has not been previously undertaken. High resolution LiDAR scanning and orthophoto imaging enabled the production of a reliable landslide inventory, but also enabled research on landslide properties and the morphology of the area. Field mapping and prospection, sampling and borehole coring assisted in the collection of information about the material characteristics and specific features of typical landslides. In the research area, which covers more than 71 km2, more than 1200 very small landslides were detected. The majority of landslides were discovered in just several geological units indicating their high susceptibility: Pleistocene silts and sands with clayey interlayers, followed by M2 silty sands and gravels, and M7 sands. Nearly half of the landslides are estimated to be of recent and younger age, while other landslides may be considered as being historical implying a “long tradition” of landslide events in the research area. Preliminary terrain surface roughness analysis also supported the conclusion that the inventory contains landslides of several historical generations which are still detectable. In addition to slides (1123), this research also discovered numerous earthflow processes (143), which are more frequent in the predominantly sandy units. The landslides in this area are largely located on the banks of the gullies and are directly related to the action of water. Regarding that situation and the engineering properties of the encountered geological units, four types of bank instabilities can be differentiated: slides on top of rock masses; slides in firm soil mixtures; landslides in sands; landslides in predominantly coherent soil complexes.

The Detection of Active Sinkholes by Airborne Differential LiDAR DEMs and InSAR Cloud Computing Tools

InSAR (Interferometric Synthetic Aperture Radar) cloud computing and the subtraction of LiDAR (Light Detection and Ranging) DEMs (Digital Elevation Models) are innovative approaches to detect subsidence in karst areas. InSAR cloud computing allows for analyzing C-band Envisat and Sentinel S1 SAR images through web platforms to produce displacement maps of the Earth’s surface in an easy manner. The subtraction of serial LiDAR DEMs results in the same product but with a different level of accuracy and precision than InSAR maps. Here, we analyze the capability of these products to detect active sinkholes in the mantled evaporite karst of the Ebro Valley (NE Spain). We found that the capability of the displacement maps produced with open access, high-resolution airborne LiDAR DEMs was up to four times higher than InSAR displacement maps generated by the Geohazard Exploitation Platform (GEP). Differential LiDAR maps provide accurate information about the location, active sectors, maximum subsidence rate and growing trend of the most rapid and damaging sinkholes. Unfortunately, artifacts and the subsidence detection limit established at −4 cm/yr entailed important limitations in the precise mapping of the sinkhole edges and the detection of slow-moving sinkholes and small collapses. Although InSAR maps provided by GEP show a worse performance when identifying active sinkholes, in some cases they can serve as a complementary technique to overcome LiDAR limitations in urban areas.

Remote sensing techniques: mapping and monitoring of mangrove ecosystem\u2014a review

Mangrove forests are considered to be the most productive ecosystem yet vanishing rapidly over the world. They are mostly found in the intertidal zone and sheltered by the seacoast. Mangroves have potential socio-economic benefits such as protecting the shoreline from storm and soil erosion, flood and flow control, acting as a carbon sink, provides a fertile breeding ground for marine species and fauna. It also acts as a source of income by providing various forest products. Restoration and conservation of mangrove forests remain a big challenge due to the large and inaccessible areas covered by mangroves forests which makes field assessment difficult and time-consuming. Remote sensing along with various digital image classification approaches seem to be promising in providing better and accurate results in mapping and monitoring the mangroves ecosystem. This review paper aims to provide a comprehensive summary of the work undertaken, and addresses various remote sensing techniques applied for mapping and monitoring of the mangrove ecosystem, and summarize their potential and limitation. For that various digital image classification techniques are analyzed and compared based on the type of image used with its spectral resolution, spatial resolution, and other related image features along with the accuracy of the classification to derive specific class information related to mangroves. The digital image classification techniques used for mangrove mapping and monitoring in various studies can be classified into pixel-based, object-based, and knowledge-based classifiers. The various satellite image data analyzed are ranged from light detection and ranging (LiDAR), hyperspectral and multispectral optical imagery, synthetic aperture radar (SAR), and aerial imagery. Supervised state of the art machine learning/deep machine learning algorithms which use both pixel-based and object-based approaches and can be combined with the knowledge-based approach are widely used for classification purpose, due to the recent development and evolution in these techniques. There is a huge future scope to study the performance of these classification techniques in combination with various high spatial and spectral resolution optical imageries, SAR and LiDAR, and also with multi-sensor, multiresolution, and temporal data.

The distribution of glacial meltwater routes and associated murtoo fields in Finland

High-resolution LiDAR (Light detection and Ranging) -based digital elevation models (DEM) have greatly improved the mapping of glacial landforms and revealed new ones such as murtoos. Murtoos have extensive diversity in form, relief and size, and they often appear along meltwater routes. However, not all meltwater routes in the recently glaciated terrains include murtoos. We mapped different types of subglacial meltwater routes and the related distribution of murtoos in the Finnish part of Fennoscandian Ice Sheet (FIS). Subglacial meltwater routes represent a previously unknown extension of the subglacial hydrological system that supplements esker networks. Murtoo deposition along the routes is dictated by the marked concentration and routing of subglacial meltwater in a high-pressure environment outside the subglacial tunnel flow and channelized drainage zone. The main environments of murtoo route genesis include the margins of glacial lineation fields, lateral shear margins of ice streams or in between ice-flow sectors or corridors, confluence zones of ice stream onset areas, lee-sides of bedrock protrusions or thresholds, bedrock fracture valleys and potential subglacial lake inputs and outputs. This study adds to the knowledge of the development of subglacial drainage and emphasizes the role of murtoos as the potential missing link between the channelized and distributed subglacial drainage.

The convexity of carbonate hilltops: 36Cl constraints on denudation and chemical weathering rates and implications for hillslope curvature

Abstract Carbonate hillslopes are often soil mantled and display a classic convex morphology. In this study we examine controls on carbonate hillslope denudation and morphology using a modified regolith mass balance equation to account for chemical weathering and dust input—two fluxes that are commonly neglected in settings with silicate-dominated bedrock. We utilize seven study sites in the Eastern Mediterranean across a significant gradient in the mean annual rainfall and dust deposition flux. Combining cosmogenic 36Cl-derived hilltop denudation rates with an estimate of the regolith chemical depletion and the quantified fraction of dust in the regolith we predict hilltop curvature and compare our predictions with observations based on high-resolution airborne LiDAR (light detection and ranging). Denudation rates vary from 5 to 210 mm/k.y. and increase with mean annual rainfall. Less resistant carbonates (chalk) experience faster denudation rates relative to more resistant dolo-limestone and are less prone to chemical weathering. Soil production exhibits a humped dependency on soil thickness. The observed hilltop curvature varies as a function of rainfall and dust flux with a minimum at sub-humid sites. While trends in hilltop convexity are often solely attributed to variations in erosion rate, our results illustrate the additional effects of dust production and chemical depletion. Our mass balance model implies that drier sites in the south probably experienced a more intricate history of regolith production due to dust flux fluctuations. Thus, by incorporating dust flux and chemical weathering to the classic hillslope evolution model we are able to identify a complex relation between hilltop curvature, soil production, and climate.

Design of an Unmanned Ground Vehicle and LiDAR Pipeline for the High-Throughput Phenotyping of Biomass in Perennial Ryegrass

Perennial ryegrass biomass yield is an important driver of profitability for Australian dairy farmers, making it a primary goal for plant breeders. However, measuring and selecting cultivars for higher biomass yield is a major bottleneck in breeding, requiring conventional methods that may be imprecise, laborious, and/or destructive. For forage breeding programs to adopt phenomic technologies for biomass estimation, there exists the need to develop, integrate, and validate sensor-based data collection that is aligned with the growth characteristics of plants, plot design and size, and repeated measurements across the growing season to reduce the time and cost associated with the labor involved in data collection. A fully automated phenotyping platform (DairyBioBot) utilizing an unmanned ground vehicle (UGV) equipped with a ground-based Light Detection and Ranging (LiDAR) sensor and Real-Time Kinematic (RTK) positioning system was developed for the accurate and efficient measurement of plant volume as a proxy for biomass in large-scale perennial ryegrass field trials. The field data were collected from a perennial ryegrass row trial of 18 experimental varieties in 160 plots (three rows per plot). DairyBioBot utilized mission planning software to autonomously capture high-resolution LiDAR data and Global Positioning System (GPS) recordings. A custom developed data processing pipeline was used to generate a plant volume estimate from LiDAR data connected to GPS coordinates. A high correlation between LiDAR plant volume and biomass on a Fresh Mass (FM) basis was observed with the coefficient of determination of R2 = 0.71 at the row level and R2 = 0.73 at the plot level. This indicated that LiDAR plant volume is strongly correlated with biomass and therefore the DairyBioBot demonstrates the utility of an autonomous platform to estimate in-field biomass for perennial ryegrass. It is likely that no single platform will be optimal to measure plant biomass from landscape to plant scales; the development and application of autonomous ground-based platforms is of greatest benefit to forage breeding programs.

A Hybrid Attention-Aware Fusion Network (HAFNet) for Building Extraction from High-Resolution Imagery and LiDAR Data

Automated extraction of buildings from earth observation (EO) data has long been a fundamental but challenging research topic. Combining data from different modalities (e.g., high-resolution imagery (HRI) and light detection and ranging (LiDAR) data) has shown great potential in building extraction. Recent studies have examined the role that deep learning (DL) could play in both multimodal data fusion and urban object extraction. However, DL-based multimodal fusion networks may encounter the following limitations: (1) the individual modal and cross-modal features, which we consider both useful and important for final prediction, cannot be sufficiently learned and utilized and (2) the multimodal features are fused by a simple summation or concatenation, which appears ambiguous in selecting cross-modal complementary information. In this paper, we address these two limitations by proposing a hybrid attention-aware fusion network (HAFNet) for building extraction. It consists of RGB-specific, digital surface model (DSM)-specific, and cross-modal streams to sufficiently learn and utilize both individual modal and cross-modal features. Furthermore, an attention-aware multimodal fusion block (Att-MFBlock) was introduced to overcome the fusion problem by adaptively selecting and combining complementary features from each modality. Extensive experiments conducted on two publicly available datasets demonstrated the effectiveness of the proposed HAFNet for building extraction.

Postfire Tree Structure from High-Resolution LiDAR and RBR Sentinel 2A Fire Severity Metrics in a Pinus halepensis-Dominated Burned Stand

Tree and plant structures remaining after fires reflect well their degree of consumption, and are therefore good indicators of fire severity. Satellite optical images are commonly used to estimate fire severity. However, depending on the severity of a fire, these sensors have a limited ability to penetrate the canopy down to the ground. Airborne light detection and ranging (LiDAR) can overcome this limitation. Assessing the differences between areas that have been burned in different fire severities based on satellite images of plant and tree structures remaining after fires is important, given its widespread use to characterize fires and fire impacts (e.g., carbon emissions). Here, we measured the remaining tree structures after a fire in a forest stand burned in SE Spain in the summer of 2017. We used high-resolution LiDAR data, acquired from an unmanned aerial vehicle (UAV) six months after the fire. This information was crossed with fire severity levels based on the relativized burnt ratio (RBR) derived from Sentinel 2A images acquired a few months before and after fire. LiDAR tree structure data derived from vertical canopy profiles (VCPs) were classified into three clusters, using hierarchical principal component analysis (HPCA), followed by a random forest (RF) to select the most important variables in distinguishing the cluster groups. Among these, crown leaf area index (LAI), crown leaf area density (LAD), crown volume, tree height and tree height skewness, among others, were the most significant variables, and reflected well the degree of combustion undergone by the trees based on the response of these variables to variations in fire severity from RBR Sentinel 2A. LiDAR metrics were able to distinguish crown fire from surface fire through changes in the understory LAI and understory and midstory vegetation. The three tree structure clusters were well separated among each other and significantly related with the RBR Sentinel 2A-derived fire severity categories. Unburned and low-severity burned areas were more diverse in tree structures than moderate and high severity burned ones. The LiDAR metrics derived from VCPs demonstrated promising potential for characterizing fine-grained post-fire plant structures and fire damage when crossed with satellite-based fire severity metrics, turning into a promising approach for better characterizing fire impacts at a resolution needed for many ecological processes.