ASG-YOLOv5: Improved YOLOv5 unmanned aerial vehicle remote sensing aerial images scenario for small object detection based on attention and spatial gating.

With the accelerated development of the technological power of society, aerial images of drones gradually penetrated various industries. Due to the variable speed of drones, the captured images are shadowed, blurred, and obscured. Second, drones fly at varying altitudes, leading to changing target scales and making it difficult to detect and identify small targets. In order to solve the above problems, an improved ASG-YOLOv5 model is proposed in this paper. Firstly, this research proposes a dynamic contextual attention module, which uses feature scores to dynamically assign feature weights and output feature information through channel dimensions to improve the model's attention to small target feature information and increase the network's ability to extract contextual information; secondly, this research designs a spatial gating filtering multi-directional weighted fusion module, which uses spatial filtering and weighted bidirectional fusion in the multi-scale fusion stage to improve the characterization of weak targets, reduce the interference of redundant information, and better adapt to the detection of weak targets in images under unmanned aerial vehicle remote sensing aerial photography; meanwhile, using Normalized Wasserstein Distance and CIoU regression loss function, the similarity metric value of the regression frame is obtained by modeling the Gaussian distribution of the regression frame, which increases the smoothing of the positional difference of the small targets and solves the problem that the positional deviation of the small targets is very sensitive, so that the model's detection accuracy of the small targets is effectively improved. This paper trains and tests the model on the VisDrone2021 and AI-TOD datasets. This study used the NWPU-RESISC dataset for visual detection validation. The experimental results show that ASG-YOLOv5 has a better detection effect in unmanned aerial vehicle remote sensing aerial images, and the frames per second (FPS) reaches 86, which meets the requirement of real-time small target detection, and it can be better adapted to the detection of the weak and small targets in the aerial image dataset, and ASG-YOLOv5 outperforms many existing target detection methods, and its detection accuracy reaches 21.1% mAP value. The mAP values are improved by 2.9% and 1.4%, respectively, compared with the YOLOv5 model. The project is available at https://github.com/woaini-shw/asg-yolov5.git.

Secure UAV adhoc network with blockchain technology.

Recent advances in aerial robotics and wireless transceivers have generated an enormous interest in networks constituted by multiple compact unmanned aerial vehicles (UAVs). UAV adhoc networks, i.e., aerial networks with dynamic topology and no centralized control, are found suitable for a unique set of applications, yet their operation is vulnerable to cyberattacks. In many applications, such as IoT networks or emergency failover networks, UAVs augment and provide support to the sensor nodes or mobile nodes in the ground network in data acquisition and also improve the overall network performance. In this situation, ensuring the security of the adhoc UAV network and the integrity of data is paramount to accomplishing network mission objectives. In this paper, we propose a novel approach to secure UAV adhoc networks, referred to as the blockchain-assisted security framework (BCSF). We demonstrate that the proposed system provides security without sacrificing the performance of the network through blockchain technology adopted to the priority of the message to be communicated over the adhoc UAV network. Theoretical analysis for computing average latency is performed based on queuing theory models followed by an evaluation of the proposed BCSF approach through simulations that establish the superior performance of the proposed methodology in terms of transaction delay, data secrecy, data recovery, and energy efficiency.

Identifying pests in precision agriculture using low-cost image data acquisition.

Unmanned Aerial Vehicles (UAVs), often called drones, have gained progressive prevalence for their swift operational ability as well as their extensive applicability in diverse real-world situations. Of late, UAV usage in precision agriculture has attracted much interest from scientific community. This study will look at drone aid in precise farming. Big data has the ability to analyze enormous amounts of data. Due to this, it is one of the diverse crucial technologies of Information and Communication Technology (ICT) which had applied in precision agriculture for the abstraction of critical information as well as for assisting agricultural practitioners in the comprehension of the most feasible farming practices, and also for better decision-making. This work analyses communication protocols, as well as their application toward the challenge of commanding a drone fleet for protecting crops from infestations of parasites. For computer-vision tasks as well as data-intensive applications, the method of deep learning has shown much potential. Due to its vast potential, it can also be used in the field of agriculture. This research will employ several schemes to assess the efficacy of models includes Visual Geometry Group (VGG-16), the Convolutional Neural Network (CNN) as well as the Fully-Convolutional Network (FCN) in plant disease detection. The methods of Artificial Immune Systems (AIS) can be used in order to adapt deep neural networks to the immediate situation. Simulated outcomes demonstrate that the proposed method is providing superior performance over various other technologically-advanced methods.

High-precision tracking and positioning for monitoring Holstein cattle.

Enhanced animal welfare has emerged as a pivotal element in contemporary precision animal husbandry, with bovine monitoring constituting a significant facet of precision agriculture. The evolution of intelligent agriculture in recent years has significantly facilitated the integration of drone flight monitoring tools and innovative systems, leveraging deep learning to interpret bovine behavior. Smart drones, outfitted with monitoring systems, have evolved into viable solutions for wildlife protection and monitoring as well as animal husbandry. Nevertheless, challenges arise under actual and multifaceted ranch conditions, where scale alterations, unpredictable movements, and occlusions invariably influence the accurate tracking of unmanned aerial vehicles (UAVs). To address these challenges, this manuscript proposes a tracking algorithm based on deep learning, adhering to the Joint Detection Tracking (JDT) paradigm established by the CenterTrack algorithm. This algorithm is designed to satisfy the requirements of multi-objective tracking in intricate practical scenarios. In comparison with several preeminent tracking algorithms, the proposed Multi-Object Tracking (MOT) algorithm demonstrates superior performance in Multiple Object Tracking Accuracy (MOTA), Multiple Object Tracking Precision (MOTP), and IDF1. Additionally, it exhibits enhanced efficiency in managing Identity Switches (ID), False Positives (FP), and False Negatives (FN). This algorithm proficiently mitigates the inherent challenges of MOT in complex, livestock-dense scenarios.

Assessing methane emissions from paddy fields through environmental and UAV remote sensing variables.

Concerns about methane (CH4) emissions from rice, a staple sustaining over 3.5 billion people globally, are heightened due to its status as the second-largest contributor to greenhouse gases, driving climate change. Accurate quantification of CH4 emissions from rice fields is crucial for understanding gas concentrations. Leveraging technological advancements, we present a groundbreaking solution that integrates machine learning and remote sensing data, challenging traditional closed chamber methods. To achieve this, our methodology involves extensive data collection using drones equipped with a Micasense Altum camera and ground sensors, effectively reducing reliance on labor-intensive and costly field sampling. In this experimental project, our research delves into the intricate relationship between environmental variables, such as soil conditions and weather patterns, and CH4 emissions. We achieved remarkable results by utilizing unmanned aerial vehicles (UAV) and evaluating over 20 regression models, emphasizing an R2 value of 0.98 and 0.95 for the training and testing data, respectively. This outcome designates the random forest regressor as the most suitable model with superior predictive capabilities. Notably, phosphorus, GRVI median, and cumulative soil and water temperature emerged as the model's fittest variables for predicting these values. Our findings underscore an innovative, cost-effective, and efficient alternative for quantifying CH4 emissions, marking a significant advancement in the technology-driven approach to evaluating rice growth parameters and vegetation indices, providing valuable insights for advancing gas emissions studies in rice paddies.

UAV-based modelling of vegetation recovery under extreme habitat stresses in the water level fluctuation zone of the Three Gorges Reservoir, China.

Impoundment of the Three Gorges Reservoir on the upper Yangtze River has remarkably altered hydrological regime within the dammed reaches, triggering structural and functional changes of the riparian ecosystem. Up to date, how vegetation recovers in response to compound habitat stresses in the water level fluctuation zone remains inexplicitly understood. In this study, plant above-ground biomass (AGB) in a selected water level fluctuation zone was quantified to depict its spatial and temporal pattern using unmanned aerial vehicle (UAV)-derived multispectral images and screened empirical models. The contributions of multiple habitat stressors in governing vegetation recovery dynamics along the environmental gradient were further explored. Screened random forest models indicated relatively higher accuracy in AGB estimation, with R2 being 0.68, 0.79 and 0.62 during the sprouting, growth, and mature periods, respectively. AGB displayed a significant linear increasing trend along the elevational gradient during the sprouting and early growth period, while it showed an inverted U-shaped pattern during late growth and mature period. Flooding duration, magnitude and timing were found to exert greater negative effects on plant sprouting and biomass accumulation and acted as decisive factors in governing the elevation-dependent pattern of AGB. Localized spatial variations in AGB were modulated by other stressors such as sediment burial, soil erosion, soil moisture and nutrient content. Occurrence of episodic summer floods and vegetation distribution were responsible for an inverted U-shaped pattern of AGB during the late growth and mature period. Generally, AGB reached its peak in August, thereafter an obvious decline by an unprecedent dry-hot climatic event. The water level fluctuations with cumulative flooding effects exerted substantial control on AGB temporal dynamics, while climatic condition played a secondary role. Herein, further restorative efforts need to be directed to screening suitable species, maintaining favorable soil condition, and improving vegetation pattern to balance the many trade-offs.

Research on Human Posture Estimation Algorithm Based on YOLO-Pose.

In response to the numerous challenges faced by traditional human pose recognition methods in practical applications, such as dense targets, severe edge occlusion, limited application scenarios, complex backgrounds, and poor recognition accuracy when targets are occluded, this paper proposes a YOLO-Pose algorithm for human pose estimation. The specific improvements are divided into four parts. Firstly, in the Backbone section of the YOLO-Pose model, lightweight GhostNet modules are introduced to reduce the model's parameter count and computational requirements, making it suitable for deployment on unmanned aerial vehicles (UAVs). Secondly, the ACmix attention mechanism is integrated into the Neck section to improve detection speed during object judgment and localization. Furthermore, in the Head section, key points are optimized using coordinate attention mechanisms, significantly enhancing key point localization accuracy. Lastly, the paper improves the loss function and confidence function to enhance the model's robustness. Experimental results demonstrate that the improved model achieves a 95.58% improvement in mAP50 and a 69.54% improvement in mAP50-95 compared to the original model, with a reduction of 14.6 M parameters. The model achieves a detection speed of 19.9 ms per image, optimized by 30% and 39.5% compared to the original model. Comparisons with other algorithms such as Faster R-CNN, SSD, YOLOv4, and YOLOv7 demonstrate varying degrees of performance improvement.

Energy-efficient UAV communication: A NOMA scheme with resource allocation and trajectory optimization.

This work investigates a downlink nonorthogonal multiple access (NOMA) scheme with unmanned aerial vehicle (UAV) aided wireless communication, where a single UAV was regarded as an air base station (ABS) to communicate with multiple ground users. Considering the constraints of velocity and maneuverability, a UAV energy efficiency (EE) model was proposed via collaborative design resource allocation and trajectory optimization. Based on this, an EE maximization problem was formulated to jointly optimize the transmit power of ground users and the trajectory of the UAV. To obtain the optimal solutions, this nonconvex problem was transformed into an equivalent convex optimization problem on the basis of three user clustering algorithms. After several alternating iterations, our proposed algorithms converged quickly. The simulation results show an enhancement in EE with NOMA because our proposed algorithm is nearly 99.6% superior to other OMA schemes.

AI-PUCMDL: artificial intelligence assisted plant counting through unmanned aerial vehicles in India's mountainous regions.

This work introduces a novel approach to remotely count and monitor potato plants in high-altitude regions of India using an unmanned aerial vehicle (UAV) and an artificial intelligence (AI)-based deep learning (DL) network. The proposed methodology involves the use of a self-created AI model called PlantSegNet, which is based on VGG-16 and U-Net architectures, to analyze aerial RGB images captured by a UAV. To evaluate the proposed approach, a self-created dataset of aerial images from different planting blocks is used to train and test the PlantSegNet model. The experimental results demonstrate the effectiveness and validity of the proposed method in challenging environmental conditions. The proposed approach achieves pixel accuracy of 98.65%, a loss of 0.004, an Intersection over Union (IoU) of 0.95, and an F1-Score of 0.94. Comparing the proposed model with existing models, such as Mask-RCNN and U-NET, demonstrates that PlantSegNet outperforms both models in terms of performance parameters. The proposed methodology provides a reliable solution for remote crop counting in challenging terrain, which can be beneficial for farmers in the Himalayan regions of India. The methods and results presented in this paper offer a promising foundation for the development of advanced decision support systems for planning planting operations.

Heatwave-like events affect drone production and brood-care behaviour in bumblebees.

Climate change is currently considered one of the major threats to biodiversity and is associated with an increase in the frequency and intensity of extreme weather events, such as heatwaves. Heatwaves create acutely stressful conditions that may lead to disruption in the performance and survival of ecologically and economically important organisms, such as insect pollinators. In this study, we investigated the impact of simulated heatwaves on the performance of queenless microcolonies of Bombus terrestris audax under laboratory conditions. Our results indicate that heatwaves can have significant impacts on bumblebee performance. However, contrary to our expectations, exposure to heatwaves did not affect survival. Exposure to a mild 5-day heatwave (30-32 °C) resulted in increased offspring production compared to those exposed to an extreme heatwave (34-36 °C) and to the control group (24 °C). We also found that brood-care behaviours were impacted by the magnitude of the heatwave. Wing fanning occurred occasionally at temperatures of 30-32 °C, whereas at 34-36 °C the proportion of workers engaged in this thermoregulatory behaviour increased significantly. Our results provide insights into the effects of heatwaves on bumblebee colony performance and underscore the use of microcolonies as a valuable tool for studying the effects of extreme weather events. Future research, especially field-based studies replicating natural foraging conditions, is crucial to complement laboratory-based studies to comprehend how heatwaves compromise the performance of pollinators. Such studies may potentially help to identify those species more resilient to climate change, as well as those that are most vulnerable.

Viability of unmanned aerial vehicles in identifying potential breeding sites for mosquito.

INTRODUCTION: Surveillance of mosquito breeding sites is essential because it provides the information needed to assess risks and thus respond to dengue outbreaks. This article aims to review existing research on the viability of using unmanned aerial vehicles (drones) to identify potential breeding sites for Aedes mosquitoes and highlight the issues related to their implementation. MATERIALS AND METHODS: The authors conducted a literature search in four databases (Scopus, Web of Science, Science Direct, and IEEE Xplore) and completed it in December 2022. Articles that do not directly address the application of drones for surveillance and control of mosquito breeding sites were excluded. RESULTS: The initial search using the keywords yielded 623 documents. After screening abstracts and reviewing the full text, only 17 articles met the inclusion criteria. Most of the studies were in the proof-of-concept stage. Many studies have also incorporated drone technologies and machine learning techniques into surveillance efforts. The authors have highlighted seven key issues related to the operational aspects of using drones. Those are hardware, software, law and regulation, operating time, expertise, geography, and community involvement. CONCLUSION: With rapid developments in drone technologies and machine learning techniques, the viability of drones as surveillance tools can be enhanced, thus effectively responding to global public health concerns.

Identification of particle distribution pattern in vertical profile via unmanned aerial vehicles observation.

Below the boundary layer, the air pollutants have been confirmed to present the decreasing trend with the height in most situaitons. However, the disperiosn rate of air pollutants in the vertical profile is rarely investigated in detail, especially through in-situ measurement. With this consideration, we employed an unmanned aerial vehicle equipped with portable monitoring equipments to scrutinize the vertical distribution of PM2.5. Based on the original data, we found that PM2.5 concentration decreases gradually with altitude below the boundary layer and demonstrated an obvious linear correlation. Therefore, the vertical distribution of PM2.5 was quantified by representing the distribution of PM2.5 with the slope of PM2.5 vertical distribution. We used backward trajectories to reveal the causes of outliers (PM2.5 increasing with altitude), and found that PM2.5 in the high altitude came from the southwest. Besides, the relationship between the vertical distribution of PM2.5 and various meteorological factors was investigated using stepwise regression analysis. The results show that the four meteorological factors most strongly correlated with the slope values are: (a) the difference in relative humidity between the ground and the air; (b) the difference in temperature between the ground and the air; (c) the height of the boundary layer; and (d) the wind speed. The slope values increase with increasing the difference in relative humidity between ground and air and the difference in temperature between the ground and the air, and decrease with increasing boundary layer height and wind speed. According to the Random Forest calculations, the ground-to-air relative humidity difference is the most important at 0.718; the wind speed is the least important at 0.053; and the ground-to-air temperature difference and boundary layer height are 0.140 and 0.088, respectively.

Short report: unmanned aerial vehicle for wide area larvicide spraying (WALS) using Vectobac® WG at Kota Kinabalu, Sabah.

INTRODUCTION: Given the stagnating progress in the fight against dengue in Kota Kinabalu, there is an urgent need to use other strategies to complement the existing vector control, focusing on larviciding. Unmanned aerial vehicle (UAV) technology has been used in vector control programs in many countries. The aim of this study was to determine the feasibility of using UAVs for larviciding to control Aedes mosquitoes in urban areas. METHODOLOGY: The Hexarotor Agro Drone (Polardrone Sdn Bhd, Malaysia) was used to carry out larviciding using Vectobac® manufactured by Valent BioSciences LLC, Libertyville, USA. The drone flew at a height of 10 feet and at a speed of 5 m/s. A total of 32 items with 10 larvae in each item were placed to test the effectiveness of larviciding using UAV. RESULTS: Out of 32 items used, 4 containers had a 100% larva mortality (13.3% mortality). The drone was not able to reach the designated spraying route that had been pre-programmed. This was due to interference with the electromagnetic waves emitted from the home satellite dishes, that were in front of the houses along the route. CONCLUSIONS: This trial showed that UAVs will be more suitable for use in larviciding in an open area without electromagnetic disturbances from electric house poles and satellite TV dishes that are commonly present in urban areas.

Incorporation of Drone Technology Into the Chain of Survival for OHCA: Estimation of Time Needed for Bystander Treatment of OHCA and CPR Performance.

BACKGROUND: Drone-delivered automated external defibrillators (AEDs) hold promises in the treatment of out-of-hospital cardiac arrest. Our objective was to estimate the time needed to perform resuscitation with a drone-delivered AED and to measure cardiopulmonary resuscitation (CPR) quality. METHODS: Mock out-of-hospital cardiac arrest simulations that included a 9-1-1 call, CPR, and drone-delivered AED were conducted. Each simulation was timed and video-recorded. CPR performance metrics were recorded by a Laerdal Resusci Anne Quality Feedback System. Multivariable regression modeling examined factors associated with time from 9-1-1 call to AED shock and CPR quality metrics (compression rate, depth, recoil, and chest compression fraction). Comparisons were made among those with recent CPR training (≤2 years) versus no recent (>2 years) or prior CPR training. RESULTS: We recruited 51 research participants between September 2019 and March 2020. The median age was 34 (Q1-Q3, 23-54) years, 56.9% were female, and 41.2% had recent CPR training. The median time from 9-1-1 call to initiation of CPR was 1:19 (Q1-Q3, 1:06-1:26) minutes. A median time of 1:59 (Q1-Q3, 01:50-02:20) minutes was needed to retrieve a drone-delivered AED and deliver a shock. The median CPR compression rate was 115 (Q1-Q3, 109-124) beats per minute, the correct compression depth percentage was 92% (Q1-Q3, 25-98), and the chest compression fraction was 46.7% (Q1-Q3, 39.9%-50.6%). Recent CPR training was not associated with CPR quality or time from 9-1-1 call to AED shock. Younger age (per 10-year increase; ß, 9.97 [95% CI, 4.63-15.31] s; P<0.001) and prior experience with AED (ß, -30.0 [95% CI, -50.1 to -10.0] s; P=0.004) were associated with more rapid time from 9-1-1 call to AED shock. Prior AED use (ß, 6.71 [95% CI, 1.62-11.79]; P=0.011) was associated with improved chest compression fraction percentage. CONCLUSION: Research participants were able to rapidly retrieve an AED from a drone while largely maintaining CPR quality according to American Heart Association guidelines. Chest compression fraction was lower than expected.

Evaluating the use of unmanned aerial vehicles for spray applications in mountain Nanguo pear orchards.

BACKGROUND: Nanguo pear is a distinctive pear variety in northeast China, grown mainly in mountainous areas. Due to terrain limitations, ground-based pesticide application equipment is difficult to use. This limitation could be overcome by using unmanned aerial vehicles (UAVs) for pesticide application in Nanguo pear orchards. This study evaluated the spraying performance of two UAVs in the Nanguo pear orchards and compared them with a manually used backpack electric sprayer (BES). The study also analyzed the effect of canopy size on droplet deposition and ground loss, and evaluated two sampling methods, leaf sampling and telescopic rod sampling. RESULTS: Compared to BESs, droplet deposition is lower for UAVs, but the actual pesticide active ingredient deposition is not necessarily lower given the solution concentration. The droplet deposition varies among different UAVs due to structural differences. Under the same UAV operating parameters, droplet deposition on trees with smaller canopy sizes is typically greater than that on trees with larger canopy sizes, and the ground loss was also more severe. Although telescopic rod sampling is a quick and convenient method, it can only reflect the trend of droplet deposition, and the data error is greater compared with leaf sampling. CONCLUSION: UAVs can achieve better droplet deposition in mountainous Nanguo pear orchards and does almost no harm to the operators compared with the BES. However, canopy size needs to be considered to adjust the application volume rate. Telescopic rods can be used for qualitative analyses, but are not recommended for quantitative analyses. © 2024 Society of Chemical Industry.

Using drone-retrieved multispectral data for phenomic selection in potato breeding.

Predictive breeding approaches, like phenomic or genomic selection, have the potential to increase the selection gain for potato breeding programs which are characterized by very large numbers of entries in early stages and the availability of very few tubers per entry in these stages. The objectives of this study were to (i) explore the capabilities of phenomic prediction based on drone-derived multispectral reflectance data in potato breeding by testing different prediction scenarios on a diverse panel of tetraploid potato material from all market segments and considering a broad range of traits, (ii) compare the performance of phenomic and genomic predictions, and (iii) assess the predictive power of mixed relationship matrices utilizing weighted SNP array and multispectral reflectance data. Predictive abilities of phenomic prediction scenarios varied greatly within a range of - 0.15 and 0.88 and were strongly dependent on the environment, predicted trait, and considered prediction scenario. We observed high predictive abilities with phenomic prediction for yield (0.45), maturity (0.88), foliage development (0.73), and emergence (0.73), while all other traits achieved higher predictive ability with genomic compared to phenomic prediction. When a mixed relationship matrix was used for prediction, higher predictive abilities were observed for 20 out of 22 traits, showcasing that phenomic and genomic data contained complementary information. We see the main application of phenomic selection in potato breeding programs to allow for the use of the principle of predictive breeding in the pot seedling or single hill stage where genotyping is not recommended due to high costs.

Development of a deep learning-based surveillance system for forest fire detection and monitoring using UAV.

This study presents a surveillance system developed for early detection of forest fires. Deep learning is utilized for aerial detection of fires using images obtained from a camera mounted on a designed four-rotor Unmanned Aerial Vehicle (UAV). The object detection performance of YOLOv8 and YOLOv5 was examined for identifying forest fires, and a CNN-RCNN network was constructed to classify images as containing fire or not. Additionally, this classification approach was compared with the YOLOv8 classification. Onboard NVIDIA Jetson Nano, an embedded artificial intelligence computer, is used as hardware for real-time forest fire detection. Also, a ground station interface was developed to receive and display fire-related data. Thus, access to fire images and coordinate information was provided for targeted intervention in case of a fire. The UAV autonomously monitored the designated area and captured images continuously. Embedded deep learning algorithms on the Nano board enable the UAV to detect forest fires within its operational area. The detection methods produced the following results: 96% accuracy for YOLOv8 classification, 89% accuracy for YOLOv8n object detection, 96% accuracy for CNN-RCNN classification, and 89% accuracy for YOLOv5n object detection.

Monitoring of Antarctica's Fragile Vegetation Using Drone-Based Remote Sensing, Multispectral Imagery and AI.

Vegetation in East Antarctica, such as moss and lichen, vulnerable to the effects of climate change and ozone depletion, requires robust non-invasive methods to monitor its health condition. Despite the increasing use of unmanned aerial vehicles (UAVs) to acquire high-resolution data for vegetation analysis in Antarctic regions through artificial intelligence (AI) techniques, the use of multispectral imagery and deep learning (DL) is quite limited. This study addresses this gap with two pivotal contributions: (1) it underscores the potential of deep learning (DL) in a field with notably limited implementations for these datasets; and (2) it introduces an innovative workflow that compares the performance between two supervised machine learning (ML) classifiers: Extreme Gradient Boosting (XGBoost) and U-Net. The proposed workflow is validated by detecting and mapping moss and lichen using data collected in the highly biodiverse Antarctic Specially Protected Area (ASPA) 135, situated near Casey Station, between January and February 2023. The implemented ML models were trained against five classes: Healthy Moss, Stressed Moss, Moribund Moss, Lichen, and Non-vegetated. In the development of the U-Net model, two methods were applied: Method (1) which utilised the original labelled data as those used for XGBoost; and Method (2) which incorporated XGBoost predictions as additional input to that version of U-Net. Results indicate that XGBoost demonstrated robust performance, exceeding 85% in key metrics such as precision, recall, and F1-score. The workflow suggested enhanced accuracy in the classification outputs for U-Net, as Method 2 demonstrated a substantial increase in precision, recall and F1-score compared to Method 1, with notable improvements such as precision for Healthy Moss (Method 2: 94% vs. Method 1: 74%) and recall for Stressed Moss (Method 2: 86% vs. Method 1: 69%). These findings contribute to advancing non-invasive monitoring techniques for the delicate Antarctic ecosystems, showcasing the potential of UAVs, high-resolution multispectral imagery, and ML models in remote sensing applications.

Use of thermal drone in detection and assessment of larval mass temperature in decomposed rabbit carcasses.

Manual ground searches and cadaver dogs are traditional methods for locating remains, but they can be time- and resource-intensive, resulting in the decomposition of bodies and delay in victim identification. Therefore, thermal imaging has been proposed as a potentially useful tool for detecting remains based on their temperature. This study investigated the potential of a novel search technique of thermal drones to detect surface remains through the detection of maggot mass temperatures. Two trials were carried out at Selangor, Malaysia, each utilizing 12 healthy male Oryctolagus cuniculus European white rabbits and DJI Matrice 300 RTK drone China, equipped with a thermal camera; Zenmuse H20T to record the thermal imaging footage of the carcasses at various heights (15, 30, 60-100 m) for 14 days for each trial. Our results demonstrated that the larval masses and corresponding heat emissions were at their largest during the active decay stage; therefore, all the carcasses were observable in thermal images on day 5 and remained until day 7. Statistical analyses showed that (1) no statistically significant differences in thermal images between clothed and unclothed subjects (p > 0.05); (2) 15 m above ground level was proven to be the optimal height, as it showed the greatest contrast between the carcass heat signature and the background (p < 0.005). Our data suggested the potential window of detection of thermal signatures was detectable up to 7 days post-deposition. This could be an important guideline for the search and recovery teams for operational implementation in this tropical region.