Animal Detection and Counting from UAV Images Using Convolutional Neural Networks

Drones, Journal Year: 2023, Volume and Issue: 7(3), P. 179 - 179

Published: March 6, 2023

In the last decade, small unmanned aerial vehicles (UAVs/drones) have become increasingly popular in airborne observation of large areas for many purposes, such as monitoring agricultural areas, tracking wild animals their natural habitats, and counting livestock. Coupled with deep learning, they allow automatic image processing recognition. The aim this work was to detect count deer population northwestern Serbia from images using neural networks, a tedious process that otherwise requires lot time effort. paper, we present compare performance several state-of-the-art network architectures, trained on manually annotated set images, use it predict presence objects rest dataset. We implemented three versions You Only Look Once (YOLO) architecture Single Shot Multibox Detector (SSD) dense forest environment measured based mean average precision (mAP), precision, recall, F1 score. Moreover, also evaluated models real-time performance. results showed selected were able up 70.45% confidence score 99%. highest achieved by fourth version YOLO 86%, well recall value 75%. Its compressed slightly lower results, 83% mAP its best case, but demonstrated four times better function applied best-performing models, providing us exact distribution over all images. Yolov4 obtained an error 8.3% counting, while Yolov4-tiny mistook 12 deer, which accounted 7.1%.

Language: Английский

---

A Survey on Monitoring of Wild Animals during Fires Using Drones

Fire, Journal Year: 2022, Volume and Issue: 5(3), P. 60 - 60

Published: April 29, 2022

Forest fires occur for natural and anthropogenic reasons affect the distribution, structure, functioning of terrestrial ecosystems worldwide. Monitoring their impacts on is an essential prerequisite effectively managing this widespread environmental problem. With development information technologies, unmanned aerial vehicles (drones) are becoming increasingly important in remote monitoring environment. One main applications drone technology related to nature observation wild animals. Unmanned thought be best solution detecting forest fires. There methods wildfires using drones with fire- and/or smoke-detection equipment. This review aims study possibility large animals during It was established that order use monitor even small groups fires, effective sensing technologies critical temperature conditions required, which can provided not only by sensors used, but also adapted software image recognition.

Language: Английский

---

Assessment of Fish Biomass and Distribution in a Nuclear Power Plant’s Water Intake Zone Using Acoustic and Trawl Methods

Animals, Journal Year: 2025, Volume and Issue: 15(7), P. 987 - 987

Published: March 29, 2025

The cooling water intake systems of coastal nuclear power plants are frequently clogged by marine organisms in the area. This study uses hydroacoustic (BioSonics DT-X, 199 kHz) and trawling methods (20 mm codend mesh size) to research fish spatiotemporal dynamics. Species composition, seasonal variations abundance, target strength were analyzed. Pearson's correlation analysis revealed a significant negative between depth resource density. Significant differences abundance biomass density observed among seasons stations. An dominant species showed that IRI (index relative importance) Collichthys lucidus was highest February 2023 (5736.54), while Harpadon nehereus had November (2309.17). distribution estimates from acoustic surveys biological highly consistent, demonstrating applicability reliability assessments. Through identification as well patterns within area, valuable data support can be provided for confirmation those prone clogging source. Moreover, it lays solid foundation development relevant protection work.

Language: Английский

---

Internal transcribed spacer primer evaluation for vascular plant metabarcoding

Metabarcoding and Metagenomics, Journal Year: 2021, Volume and Issue: 5

Published: Sept. 2, 2021

The unprecedented ongoing biodiversity decline necessitates scalable means of monitoring in order to fully understand the underlying causes. DNA metabarcoding has potential provide a powerful tool for accurate and rapid monitoring. Unfortunately, many cases, lack universal standards undermines widespread application metabarcoding. One most important considerations plants, aside from selecting potent barcode marker, is primer choice. Our study evaluates published ITS primers silico vitro , through mock communities presents newly designed primers. We were able show that large proportion previously available have unfavourable attributes. combined results support recommendation introduced ITS-3p62plF1 ITS-4unR1 as best current plant specific ITS2 combination. also found PCR optimisation, such addition 5% DMSO, essential obtain meaningful Finally, we conclude continuous quality assurance indispensable reliable results.

Language: Английский

---

A framework for monitoring biodiversity in protected areas and other effective area-based conservation measures: concepts, methods and technologies

Published: April 3, 2024

Protected areas and other effective area-based conservation measures (OECMs) are important to stop the global decline in biodiversity. Systematic site-based monitoring of state biodiversity outcomes is necessary for evidence-based adaptive management protected OECMs. Biodiversity also inform managers if they meeting their goals. The framework described this publication will help site planners consider all relevant details develop programmes improved outcomes.

Language: Английский

---

Automated Detection of Koalas with Deep Learning Ensembles

Remote Sensing, Journal Year: 2022, Volume and Issue: 14(10), P. 2432 - 2432

Published: May 19, 2022

Effective management of threatened and invasive species requires regular reliable population estimates. Drones are increasingly utilised by ecologists for this purpose as they relatively inexpensive. They enable larger areas to be surveyed than traditional methods many species, particularly cryptic such koalas, with less disturbance. The development robust accurate detection is required effectively use the large volumes data generated survey method. enhanced predictive computational power deep learning ensembles represents a considerable opportunity ecological community. In study, we investigate potential built from multiple convolutional neural networks (CNNs) detect koalas low-altitude, drone-derived thermal data. approach uses detectors combinations YOLOv5 models Detectron2. achieved strong balance between probability precision when tested on ground-truth radio-collared koalas. Our results also showed that greater diversity in ensemble composition can enhance overall performance. We found main impediment higher was false positives but expect these will continue reduce tools geolocating detections improved. ability construct different sizes allow improved alignment algorithms used characteristics problems. Ensembles efficient scaled suit settings, platforms hardware availability, making them capable adaption novel applications.

Language: Английский

---

Internet of Wild Things with the Integration of Vision Technology and LoRa Network

International Journal of Electrical and Electronics Engineering, Journal Year: 2023, Volume and Issue: 10(3), P. 1 - 7

Published: March 25, 2023

Sustainable Development Goals (SDGs) are significant in protecting the habitat of wild animals and preventing animal fatalities resulting from human development. The construction land transportation systems through wildlife has compromised life animals, thus invoking implementation advanced technologies such as Internet Things (IoT). Motivated by these aspects, this article proposes a system incorporating IoT LoRa-based technology to prevent human-wildlife confrontations creating real-time warning commuters about movement on road. proposed system, consisting vision node, display layer data visualization layer, was realized real time deploying selected road corridor. Vision node is transmitted LoRa communication visualized serial monitor.

Language: Английский

---

Case study of detection and monitoring of wildlife by UAVs equipped with RGB camera and TIR camera

Aircraft Engineering and Aerospace Technology, Journal Year: 2023, Volume and Issue: 95(10), P. 1461 - 1469

Published: Sept. 21, 2023

Purpose Observation of the animal world is an important component nature surveys. It provides a number different information concerning aspects such as population sizes, migration directions, feeding sites and many other data. The paper below presents results from flights unmanned aerial vehicle (UAV) aimed at detecting animals in their natural environment. Design/methodology/approach drone used research was equipped with RGB thermal infrared (TIR) cameras. Both cameras, which were mounted on UAV, to take pictures showing concentration (deer). overview carried out villages Podlaskie Voivodeship: Szerokie Laki, Bialousy Sloja. Research made A deer photographed during Durango platform, imaging camera Canon camera, for flights. taken create orthomaps. multicopter, GoPro film flight locations. control station also used, consisting laptop MissionPlanner software. Findings Analysis collected images has indicated that environmental, organisational technical factors influence quality information. Sophisticated observation precision ensured by use high-resolution TIR proper platform cameras UAV provided advanced positioning systems, makes it possible high-quality orthomaps area. When observing animals, time day (temperature contrast), year season (leaf ascent) or parameters important. Originality/value introduces conclusions flights, pointing useful using UAVs.

Language: Английский

---

LAND REMOTE SENSING TECHNOLOGIES USING UNMANNED AERIAL VEHICLES

Успехи современного естествознания (Advances in Current Natural Sciences), Journal Year: 2024, Volume and Issue: №1 2024, P. 74 - 79

Published: Jan. 1, 2024

В последние годы возросло применение дистанционного зондирования беспилотных летательных аппаратов (БПЛА) для мониторинга техногенно нарушенных земель, а направления применения расширились.И в настоящее время было проведено несколько обзорных исследований, посвященных конкретно нарушенным землям.Однако недостаточно систематических и всесторонних исследований воздушных судов мониторинге экосистем.Целью данной работы являлось изучение технологии Земли с применением контроля земель Кемеровской области -Кузбасса.Впервые проанализировали тенденции земель.Основными аспектами сценариев являются мониторинг растительности, физический химический почвы, деградации почвы нарушений окружающей среды.Существующие ограничения развития были обобщены.Приведенные материалы будут полезны лучшего понимания возможностей помощью не только угольных участков, но научных исследованиях биологических

Language: Русский

---

Hunting Resource Management by Population Size Control by Remote Sensing Using an Unmanned Aerial Vehicle

Nature Environment and Pollution Technology, Journal Year: 2024, Volume and Issue: 23(1), P. 391 - 399

Published: March 1, 2024

The study was carried out on the territory of Kemerovo region-Kuzbass (Western Siberia, Russia). purpose to obtain information species diversity and population big-game animals. monitoring forest territories region’s administrative districts. In course remote sensing using an unmanned aerial vehicle, presence all types animals under consideration, except for bear, recorded. deviation number determined traditional method digital technologies varied up 50%. It established that environmental measures organized by regional administration hunting farms improved situation stabilized main group game found when a sufficiently high sensitivity thermal imager (the used had very class ≤ 60 mK at 300 K), long-haired animals, which are characterized lower intensity radiation (for example, wolves) identified recognized in images. larger animal worse insulation layer (wool or feathers), easier it is identify infrared images requirements imagers. ability recognize record smaller birds requires additional research existing technologies. Our has confirmed validity methods managing wildlife nature conservation areas Siberian Taiga territories.

Language: Английский

---