Towards Mobile Wind Measurements Using Joust Configured Ultrasonic Anemometer for Applications in Gas Flux Quantification

Drones, Journal Year: 2025, Volume and Issue: 9(2), P. 94 - 94

Published: Jan. 26, 2025

Small uncrewed aerial systems (sUASs) can be used to quantify emissions of greenhouse and other gases, providing flexibility in quantifying these from a multitude sources, including oil gas infrastructure, volcano plumes, wildfire emissions, natural sources. However, sUAS-based emission estimates are sensitive the accuracy wind speed direction measurements. In this study, we examined how filtering correcting measurements affects data by comparing miniature ultrasonic anemometer mounted on sUAS joust configuration highly accurate taken nearby eddy covariance flux tower (aka Tower). These corrections had small effect error, but reduced errors 50° >120° 20–30°. A concurrent experiment examining amount error due Tower not being co-located showed that impact separation was 0.16–0.21 ms−1, influence errors. Lower were correlated with lower turbulence intensity higher relative speeds. There also some loose trends diminished at Therefore, improve quality measurements, our study suggested flight planning consider optimizing conditions maximize speeds as well include post-flight corrections.

Language: Английский

---

Modelling LiDAR-Based Vegetation Geometry for Computational Fluid Dynamics Heat Transfer Models

Remote Sensing, Journal Year: 2025, Volume and Issue: 17(3), P. 552 - 552

Published: Feb. 6, 2025

Vegetation characteristics significantly influence the impact of wildfires on individual building structures, and these effects can be systematically analyzed using heat transfer modelling software. Close-range light detection ranging (LiDAR) data obtained from uncrewed aerial systems (UASs) capture detailed vegetation morphology; however, integration dense merged canopies into three-dimensional (3D) models for fire software poses significant challenges. This study proposes a method integrating UAS–LiDAR-derived geometric features components—such as bark, wooden core, foliage—into models. The were collected natural woodland surrounding an elevated in Samford, Queensland, Australia. Aboveground biomass (AGB) was estimated 21 trees utilizing three 3D tree reconstruction tools, with validation against allometric equations (BAEs) derived field measurements. most accurate tool produced mesh utilized geometry. A proof concept established Eucalyptus siderophloia, incorporating non-destructive framework leverages available technologies to create reliable reconstructions complex wildland–urban interfaces (WUIs). It facilitates realistic wildfire risk assessments by providing flux estimations, which are critical evaluating safety during events, while addressing limitations associated direct

Language: Английский

---

Identification of Vegetation Areas Affected by Wildfires Using RGB Images Obtained by UAV: A Case Study in the Brazilian Cerrado

Geomatics, Journal Year: 2025, Volume and Issue: 5(1), P. 13 - 13

Published: March 16, 2025

The Cerrado is Brazil’s second largest biome, covering continuous areas in several states. Covering approximately 23% of territory, the biome connects with all main biomes South America, thus forming a major biological corridor. This one those that has suffered most from incidence wildfires, leading to progressive depletion region’s natural resources. aim this study was evaluate use an Unmanned Aerial Vehicle (UAV) embedded RGB sensor obtain high-resolution digital products can be used identify Brazilian affected by wildfires. carried out savannah area selecting vegetation corridor native free anthropogenic influence. following UAV surveys were before and after burning event. Once orthomosaics available, GLI, VARI, ExG NGRDI indices analyze vegetation. data indicate B band GLI are more suitable for environmental impact analysis fires, providing solid basis monitoring management scenarios fire disturbance.

Language: Английский

---

Characterization of Shrub Fuel Structure and Spatial Distribution Using Multispectral and 3D Multitemporal UAV Data

Forests, Journal Year: 2025, Volume and Issue: 16(4), P. 676 - 676

Published: April 12, 2025

Shrubland vegetation plays a crucial role in ecological processes, but its conservation is facing threats due to climate change, wildfires, and human activities. Unmanned Aerial Vehicles (UAVs), or ‘drones’, have become valuable tools for detailed mapping, providing high-resolution imagery 3D models despite challenges such as legal restrictions limited coverage. We developed methodology estimating height, map classes, fuel by using multitemporal UAV data (imagery point clouds from the imagery) other ancillary provide insights into habitat condition characteristics. Two different random forest classification methods (an object- pixel-based approach) discriminating between classes were compared. The method showed promise characterizing structure (shrub height), with an RMSE of less than 0.3 m slight overestimation taller heights. For models, best results obtained object-based approach, overall accuracies 0.96 0.93, respectively. Although some difficulties encountered distinguishing low shrubs brackens low-height spatial mixture, accurate most classes. Future improvements include refining terrain including acquired aerial scanners exploring phenological stages machine learning approaches classification.

Language: Английский

---

Conceptual design of a wildfire emergency response system empowered by swarms of unmanned aerial vehicles

International Journal of Disaster Risk Reduction, Journal Year: 2025, Volume and Issue: unknown, P. 105493 - 105493

Published: April 1, 2025

Language: Английский

---

Design and implementation for a UAV-based streaming media system

Ad Hoc Networks, Journal Year: 2024, Volume and Issue: 156, P. 103443 - 103443

Published: Feb. 12, 2024

Language: Английский

---

Multi-Objective Deployment of UAVs for Multi-Hop FANET: UAV-Assisted Emergency Vehicular Network

Drones, Journal Year: 2024, Volume and Issue: 8(6), P. 262 - 262

Published: June 13, 2024

In the event of a sudden natural disaster, damaged communication infrastructure cannot provide necessary network service for vehicles. Unfortunately, this is critical moment when occupants trapped vehicles need to urgently use vehicular network’s emergency service. How efficiently connect vehicle base station challenge facing network. To address challenge, study proposes UAV-assisted multi-objective and multi-hop ad hoc (UMMVN) that can be used as an Firstly, it presents integrated design search system find vehicle, relay, networking, which significantly decreases UAV’s networking time cost. Secondly, navigates UAVs along multiple paths different objectives. Thirdly, optimal branching node strategy allows adequate overlapping targets, cost within limited searching range. The numerical experiments illustrate UMMVN performs better than other state-of-the-art methods.

Language: Английский

---

African Lovegrass Segmentation with Artificial Intelligence Using UAS-Based Multispectral and Hyperspectral Imagery

Remote Sensing, Journal Year: 2024, Volume and Issue: 16(13), P. 2363 - 2363

Published: June 27, 2024

The prevalence of the invasive species African Lovegrass (Eragrostis curvula, ALG thereafter) in Australian landscapes presents significant challenges for land managers, including agricultural losses, reduced native diversity, and heightened bushfire risks. Uncrewed aerial system (UAS) remote sensing combined with AI algorithms offer a powerful tool accurately mapping spatial distribution facilitating effective management strategies. However, segmentation vegetations within mixed grassland ecosystems due to heterogeneity, spectral similarity, seasonal variability. performance state-of-the-art artificial intelligence (AI) detecting landscape remains unknown. This study compared four supervised models segmenting using multispectral (MS) imagery at sites developed two different conditions. UAS surveys were conducted New South Wales, Australia. Two surveyed distinct seasons (flowering vegetative), each comprised data collection settings. A comparative analysis was also between hyperspectral (HS) MS single site flowering season. Of five (XGBoost, RF, SVM, CNN, U-Net), XGBoost customized CNN model achieved highest validation accuracy 99%. testing used approaches: quadrat-based proportion prediction environments pixel-wise classification masked regions where other classes could be confidently differentiated. Quadrat-based ground truth values against custom model, resulting 5.77% 12.9% RMSE seasons, respectively, emphasizing superiority over algorithms. comparison U-Net demonstrated that effectively captures without requiring more intricate architecture U-Net. Masked-based results showed higher F1 scores, 91.68% season 90.61% vegetative Models trained on single-season exhibited decreased when evaluated from varying Integrating both during training resulted reduction error out-of-season predictions, suggesting improved generalizability through multi-season integration. Moreover, HS predictions similar test around 20% proportion, highlighting practicality operational limitations. shows great promise biodiversity conservation by sustainable strategies controlling spread.

Language: Английский

---

Using computer vision to detect and segment fire behavior classifications in UAS-captured images

EarthArXiv (California Digital Library), Journal Year: 2024, Volume and Issue: unknown

Published: Feb. 21, 2024

The widely adaptable capabilities of artificial intelligence, in particular deep learning and computer vision has led to significant research output regarding fire smoke detection. Previous studies often focus on themes like early detection, increased operational awareness, post-fire assessment. To further test the detection these scenarios, we collected labeled a unique aerial image dataset that determined whether specific types behavior could be reliably detected prescribed settings. Our 960 images were sourced from over 20.97 hours UAS video during operations covering large region Texas Louisiana, U.S.. National Wildfire Coordinating Group (NWCG) observations descriptions served as reference for determining classes labeling. YOLOv8 models trained NWCG Rank 1-3 grassland, shrubland, forested, combined regimes within our study area. Models first validated isolated objects behavior, then segmenting their original parent images. using consistently performed at mAP 0.808 or higher, with producing best results (mAP = 0.897). Most segmentation relatively poorly, except forest regime model box mask 0.59 0.611, respectively. indicate classifying is possible most fuel models, whereas around background information difficult. However, it may manageable task enough data, when are developed regime. With an increasing number destructive wildfires new challenges confronting managers, identifying how technologies can quickly assess wildfire situations assist responder awareness. conclusion levels abstraction deeper than mere vision, make even more detailed monitoring possible.

Language: Английский

---

Domestic experience of unmanned aerial vehicles for forest conditions monitoring

BIO Web of Conferences, Journal Year: 2024, Volume and Issue: 93, P. 01028 - 01028

Published: Jan. 1, 2024

Russia has vast forest resources that require constant conservation and protection measures, the implementation of which is currently impossible without use aviation or satellites. The widespread introduction unmanned aerial vehicles will make it possible to effectively monitor condition predict changes in lands. article discusses standard methods for monitoring forests their advantages disadvantages, highlighting main existing threats forests. classification given. using forestry are highlighted. A areas application

Language: Английский

---