Margarita Mulero-Pázmány
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Margarita Mulero-Pázmány has a total of 178 co-authors. Here are the top co-authors by number of shared publications:
- Juan Jose Negro
Doñana Biological Station, CSIC 2 Department of Ecology and Evolution Avda. Américo Vespucio s/n, 41092 Sevilla
5 shared publications Juan José Negro
4 shared publications- Manuela González-Suárez
Ecology and Evolutionary Biology, School of Biological Sciences, University of Reading, Reading, UK
4 shared publications - Marcello D'AMico
Estación Biológica de Doñana CSIC Seville Spain
4 shared publications - Serge A. Wich
School of Biological and Environmental Sciences, Liverpool John Moores University, Liverpool, UK
4 shared publications Joaquín Vicente
3 shared publications
Latest Publications
- Open Access
- 17 July 2026
- Published by Universidad de Zaragoza in Jornada de Jóvenes Investigadores del I3A
- Vol. 14
Herramienta interactiva para segmentación de imágenes con SAM2
Este trabajo presenta una herramienta para la segmentación interactiva de información semántica relevante en imágenes de entornos naturales basado en modelos fundacionales. La nueva herramienta facilita el uso del modelo SAM2, por parte de usuarios no expertos en el manejo de modelos fundacionales de visión. A modo de ejemplo, se muestra el uso de la herramienta en la segmentación de caminos en ortofotos aéreas.
- 17 July 2026
- Published by Universidad de Zaragoza in Jornada de Jóvenes Investigadores del I3A
- Vol. 14
Open AccessHerramienta interactiva para segmentación de imágenes con SAM2
Este trabajo presenta una herramienta para la segmentación interactiva de información semántica relevante en imágenes de entornos naturales basado en modelos fundacionales. La nueva herramienta facilita el uso del modelo SAM2, por parte de usuarios no expertos en el manejo de modelos fundacionales de visión. A modo de ejemplo, se muestra el uso de la herramienta en la segmentación de caminos en ortofotos aéreas.
- 26 February 2026
- Published by Springer Nature
- p. 433-441
Technological Methods in Road Ecology
- Top Cited PapersOpen Access
- 9 May 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15 (1) , 1-18
Addressing significant challenges for animal detection in camera trap images: a novel deep learning-based approach
Wildlife biologists increasingly use camera traps for monitoring animal populations. However, manually sifting through the collected images is expensive and time-consuming. Current deep learning studies for camera trap images do not adequately tackle real-world challenges such as imbalances between animal and empty images, distinguishing similar species, and the impact of backgrounds on species identification, limiting the models’ applicability in new locations. Here, we present a novel two-stage deep learning framework. First, we train a global deep-learning model using all animal species in the dataset. Then, an agglomerative clustering algorithm groups animals based on their appearance. Subsequently, we train a specialized deep-learning expert model for each animal group to detect similar features. This approach leverages Transfer Learning from the MegaDetectorV5 (YOLOv5 version) model, already pre-trained on various animal species and ecosystems. Our two-stage deep learning pipeline uses the global model to redirect images to the appropriate expert models for final classification. We validated this strategy using 1.3 million images from 91 camera traps encompassing 24 mammal species and used 120,000 images for testing, achieving an F1-Score of 96.2% using expert models for final classification. This method surpasses existing deep learning models, demonstrating improved precision and effectiveness in automated wildlife detection.
- 9 May 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15 (1) , 1-18
Top Cited PapersOpen AccessAddressing significant challenges for animal detection in camera trap images: a novel deep learning-based approach
Wildlife biologists increasingly use camera traps for monitoring animal populations. However, manually sifting through the collected images is expensive and time-consuming. Current deep learning studies for camera trap images do not adequately tackle real-world challenges such as imbalances between animal and empty images, distinguishing similar species, and the impact of backgrounds on species identification, limiting the models’ applicability in new locations. Here, we present a novel two-stage deep learning framework. First, we train a global deep-learning model using all animal species in the dataset. Then, an agglomerative clustering algorithm groups animals based on their appearance. Subsequently, we train a specialized deep-learning expert model for each animal group to detect similar features. This approach leverages Transfer Learning from the MegaDetectorV5 (YOLOv5 version) model, already pre-trained on various animal species and ecosystems. Our two-stage deep learning pipeline uses the global model to redirect images to the appropriate expert models for final classification. We validated this strategy using 1.3 million images from 91 camera traps encompassing 24 mammal species and used 120,000 images for testing, achieving an F1-Score of 96.2% using expert models for final classification. This method surpasses existing deep learning models, demonstrating improved precision and effectiveness in automated wildlife detection.
- Open Access
- 1 December 2023
- Published by The Raptor Research Foundation, Inc. in Journal of Raptor Research
- Vol. 57 (4) , 667-670
Wildlife Tourism Malpractice Can Lead to Animal Poisoning with Plastics
The Journal of Raptor Research publishes the biology and conservation of diurnal and nocturnal raptors, such as their ecology, behavior, and conservation.
- 1 December 2023
- Published by The Raptor Research Foundation, Inc. in Journal of Raptor Research
- Vol. 57 (4) , 667-670
Open AccessWildlife Tourism Malpractice Can Lead to Animal Poisoning with Plastics
The Journal of Raptor Research publishes the biology and conservation of diurnal and nocturnal raptors, such as their ecology, behavior, and conservation.
- Open Access
- 26 August 2023
- Published by Wiley in Remote Sensing in Ecology and Conservation
- Vol. 9 (4) , 435-438
Issue Information
- 26 August 2023
- Published by Wiley in Remote Sensing in Ecology and Conservation
- Vol. 9 (4) , 435-438
Open AccessIssue Information
- Open Access
- 1 June 2023
- Published by Bristol University Press in Global Social Challenges Journal
- Vol. 2 (aop) , 2-23
Exploring the opportunities and risks of aerial monitoring for biodiversity conservation
Drones are unoccupied aerial systems (UAS) whose technology has evolved rapidly over the past 15 years. Increasingly used in conservation to manage and monitor biodiversity, drones offer rich capabilities to observe in difficult terrain, have relatively affordable hardware costs and are likely to continue to proliferate rapidly in the years ahead. Drones are useful for tasks as diverse as monitoring wildlife poaching and illegal timber extraction, managing ecotourism and disaster responses, and tracking the regeneration or degradation of forests, and offer potential for more specialised tasks as their sensory payloads are developed. However, although associated technical issues and applications have been explored in wide-ranging ways within conservation science, there has been relatively little social-scientific engagement with drones to date. This leaves a gap surrounding the potential social benefits and risks of drones, as well as in interdisciplinary conversations. This introduction is the first of four papers under the heading ‘Drone ecologies’, building on an interdisciplinary workshop held under the same name at the University of Bristol in July 2021. Expanding from the plenary dialogues that opened this workshop, this introduction explores what interdisciplinary perspectives on drones can offer in addressing global social and ecological challenges, drawing on expertise from the fields of conservation biology, human and physical geography, rainforest ecology and environmental systems. Setting out the aims of the overall special collection, we review here the ways that drones are being used, and might be used, in biodiversity conservation, setting out important considerations to minimise risks of inadvertent harms.
- 1 June 2023
- Published by Bristol University Press in Global Social Challenges Journal
- Vol. 2 (aop) , 2-23
Open AccessExploring the opportunities and risks of aerial monitoring for biodiversity conservation
Drones are unoccupied aerial systems (UAS) whose technology has evolved rapidly over the past 15 years. Increasingly used in conservation to manage and monitor biodiversity, drones offer rich capabilities to observe in difficult terrain, have relatively affordable hardware costs and are likely to continue to proliferate rapidly in the years ahead. Drones are useful for tasks as diverse as monitoring wildlife poaching and illegal timber extraction, managing ecotourism and disaster responses, and tracking the regeneration or degradation of forests, and offer potential for more specialised tasks as their sensory payloads are developed. However, although associated technical issues and applications have been explored in wide-ranging ways within conservation science, there has been relatively little social-scientific engagement with drones to date. This leaves a gap surrounding the potential social benefits and risks of drones, as well as in interdisciplinary conversations. This introduction is the first of four papers under the heading ‘Drone ecologies’, building on an interdisciplinary workshop held under the same name at the University of Bristol in July 2021. Expanding from the plenary dialogues that opened this workshop, this introduction explores what interdisciplinary perspectives on drones can offer in addressing global social and ecological challenges, drawing on expertise from the fields of conservation biology, human and physical geography, rainforest ecology and environmental systems. Setting out the aims of the overall special collection, we review here the ways that drones are being used, and might be used, in biodiversity conservation, setting out important considerations to minimise risks of inadvertent harms.
- Open Access
Terrestrial Megafauna Response to Drone Noise Levels in Ex Situ Areas
Drone use has significantly grown in recent years, and there is a knowledge gap on how the noise produced by these systems may affect animals. We investigated how 18 species of megafauna reacted to drone sound pressure levels at different frequencies. The sound pressure level on the low frequency generated by the drone did not change the studied species’ behavior, except for the Asian elephant. All other studied species showed higher noise sensitivity at medium and high frequencies. The Asian elephant was the most sensitive species to drone noise, mainly at low frequencies. Felines supported the highest sound pressure level before showing behavioral reactions. Our results suggest that drone sound pressure levels in different frequencies cause behavioral changes that differ among species, which is relevant to assessing drone disturbances in ex situ environments. The findings presented here can help to reduce drone impact for target species and serve as an experimental study for future drone use guidelines.
Open AccessTerrestrial Megafauna Response to Drone Noise Levels in Ex Situ Areas
Drone use has significantly grown in recent years, and there is a knowledge gap on how the noise produced by these systems may affect animals. We investigated how 18 species of megafauna reacted to drone sound pressure levels at different frequencies. The sound pressure level on the low frequency generated by the drone did not change the studied species’ behavior, except for the Asian elephant. All other studied species showed higher noise sensitivity at medium and high frequencies. The Asian elephant was the most sensitive species to drone noise, mainly at low frequencies. Felines supported the highest sound pressure level before showing behavioral reactions. Our results suggest that drone sound pressure levels in different frequencies cause behavioral changes that differ among species, which is relevant to assessing drone disturbances in ex situ environments. The findings presented here can help to reduce drone impact for target species and serve as an experimental study for future drone use guidelines.
- Open Access
Development of a Fixed-Wing Drone System for Aerial Insect Sampling
Most current insect research techniques are ground-based and provide scarce information about flying insects in the planetary boundary layer (PBL), which remains a poorly studied ecological niche. To address this gap, we developed a new insect-sampling method consisting of a fixed-wing drone platform with net traps attached to the fuselage, a mobile design that has optimal aerodynamic characteristics for insect capture in the PBL. We tested the proposed device on 16 flights in Doñana National Park (Spain) with two different trap designs fitted on the fuselage nose and wing. We collected 34 insect specimens belonging to four orders with a representation of twelve families at mean altitudes below 23 m above ground level and sampling altitudes between 9 and 365 m. This drone insect-sampling design constitutes a low-cost and low-impact method for insect monitoring in the PBL, especially in combination with other remote sensing technologies that directly quantify aerial insect abundance but do not provide taxonomic information, opening interesting possibilities for ecology and entomological research, with the possibility of transfer to economically important sectors, such as agriculture and health.
Open AccessDevelopment of a Fixed-Wing Drone System for Aerial Insect Sampling
Most current insect research techniques are ground-based and provide scarce information about flying insects in the planetary boundary layer (PBL), which remains a poorly studied ecological niche. To address this gap, we developed a new insect-sampling method consisting of a fixed-wing drone platform with net traps attached to the fuselage, a mobile design that has optimal aerodynamic characteristics for insect capture in the PBL. We tested the proposed device on 16 flights in Doñana National Park (Spain) with two different trap designs fitted on the fuselage nose and wing. We collected 34 insect specimens belonging to four orders with a representation of twelve families at mean altitudes below 23 m above ground level and sampling altitudes between 9 and 365 m. This drone insect-sampling design constitutes a low-cost and low-impact method for insect monitoring in the PBL, especially in combination with other remote sensing technologies that directly quantify aerial insect abundance but do not provide taxonomic information, opening interesting possibilities for ecology and entomological research, with the possibility of transfer to economically important sectors, such as agriculture and health.
- Open Access
- 14 July 2022
- Published by CSIRO Publishing in Wildlife Research
- Vol. 50 (1) , 39-46
Road orientation affects the impact of roads on wildlife
Context: Understanding the factors determining the impacts of roads and how they fragment landscapes limiting the movement of animals, is key to implement efficient mitigation measures. Aims: Here we investigate if road orientation in relation to limiting resources, a largely overlooked factor on road impact assessments, can influence the movement of animals within a landscape where water resources are spatially clustered. Methods: We evaluated movement by monitoring animal tracks on unpaved roads: two with a North–South orientation and two with an East–West orientation. Key results: Animals were more likely to follow roads leading to limiting resources (i.e. East–West orientation), confirming human linear structures can facilitate wildlife movements. Carnivores were more likely to follow roads with any orientation and for longer compared to ungulates, whereas ungulates followed roads mainly in the orientation of limiting resources. Conclusions: Road orientation affects how roads influence the movement of animals in landscapes where resources are distributed along a spatial gradient with different effects for ungulates and carnivores. Implications: The key implications of this work affect the planning and implementation of mitigation strategies and safety measures. Our results suggest road-crossing infrastructure and fences will be most important in roads traversing a gradient to allow wildlife movement while preventing collisions. For roads along a gradient, crossing structures may be less important, but fences or appropriate signage could be useful to prevent or warn drivers of animals travelling on the road.
- 14 July 2022
- Published by CSIRO Publishing in Wildlife Research
- Vol. 50 (1) , 39-46
Open AccessRoad orientation affects the impact of roads on wildlife
Context: Understanding the factors determining the impacts of roads and how they fragment landscapes limiting the movement of animals, is key to implement efficient mitigation measures. Aims: Here we investigate if road orientation in relation to limiting resources, a largely overlooked factor on road impact assessments, can influence the movement of animals within a landscape where water resources are spatially clustered. Methods: We evaluated movement by monitoring animal tracks on unpaved roads: two with a North–South orientation and two with an East–West orientation. Key results: Animals were more likely to follow roads leading to limiting resources (i.e. East–West orientation), confirming human linear structures can facilitate wildlife movements. Carnivores were more likely to follow roads with any orientation and for longer compared to ungulates, whereas ungulates followed roads mainly in the orientation of limiting resources. Conclusions: Road orientation affects how roads influence the movement of animals in landscapes where resources are distributed along a spatial gradient with different effects for ungulates and carnivores. Implications: The key implications of this work affect the planning and implementation of mitigation strategies and safety measures. Our results suggest road-crossing infrastructure and fences will be most important in roads traversing a gradient to allow wildlife movement while preventing collisions. For roads along a gradient, crossing structures may be less important, but fences or appropriate signage could be useful to prevent or warn drivers of animals travelling on the road.
- Open Access
- 18 April 2022
- Published by Oxford University Press (OUP) in Current Zoology
- Vol. 69 (2) , 208-214
A practical approach with drones, smartphones, and tracking tags for potential real-time animal tracking
Drones are increasingly used for fauna monitoring and wildlife tracking; however, their application for wildlife tracking is restricted by developing such systems. Here we explore the potential of drones for wildlife tracking using an off-the-shelf system that is easy to use by non-specialists consisting of a multirotor drone, smartphones, and commercial tracking devices via Bluetooth and Ultra-Wide Band (UWB). We present the system configuration, explore the operational parameters that can affect detection capabilities, and test the effectiveness of the system for locating targets by simulating target animals in savanna and forest environments. The self-contained tracking system was built without hardware or software customization. In 40 tracking flights carried out in the Brazilian Cerrado, we obtained a detection rate of 90% in savanna and 40% in forest areas. Tests for targets in movement (N = 20), the detection rates were 90% in the savanna and 30% in the forest areas. The spatial accuracy obtained by the system was 14.61 m, being significantly more accurate in savanna ( = 10.53) than in forest areas ( = 13.06). This approach to wildlife tracking facilitates the use of drones by non-specialists at an affordable cost for conservation projects with limited resources. The reduced size of the tags, the long battery life, and the lower cost compared to GPS-tags open up a range of opportunities for animal tracking.- 18 April 2022
- Published by Oxford University Press (OUP) in Current Zoology
- Vol. 69 (2) , 208-214
Open AccessA practical approach with drones, smartphones, and tracking tags for potential real-time animal tracking
Drones are increasingly used for fauna monitoring and wildlife tracking; however, their application for wildlife tracking is restricted by developing such systems. Here we explore the potential of drones for wildlife tracking using an off-the-shelf system that is easy to use by non-specialists consisting of a multirotor drone, smartphones, and commercial tracking devices via Bluetooth and Ultra-Wide Band (UWB). We present the system configuration, explore the operational parameters that can affect detection capabilities, and test the effectiveness of the system for locating targets by simulating target animals in savanna and forest environments. The self-contained tracking system was built without hardware or software customization. In 40 tracking flights carried out in the Brazilian Cerrado, we obtained a detection rate of 90% in savanna and 40% in forest areas. Tests for targets in movement (N = 20), the detection rates were 90% in the savanna and 30% in the forest areas. The spatial accuracy obtained by the system was 14.61 m, being significantly more accurate in savanna ( = 10.53) than in forest areas ( = 13.06). This approach to wildlife tracking facilitates the use of drones by non-specialists at an affordable cost for conservation projects with limited resources. The reduced size of the tags, the long battery life, and the lower cost compared to GPS-tags open up a range of opportunities for animal tracking.