Counting Mixed Breeding Aggregations of Animal Species Using Drones: Lessons from Waterbirds on Semi-Automation

Optimized Small Waterbird Detection Method Using Surveillance Videos Based on YOLOv7

Waterbird monitoring is the foundation of conservation and management strategies in almost all types of wetland ecosystems. China's improved wetland protection infrastructure, which includes remote devices for the collection of larger quantities of acoustic and visual data on wildlife species, increased the need for data filtration and analysis techniques. Object detection based on deep learning has emerged as a basic solution for big data analysis that has been tested in several application fields. However, these deep learning techniques have not yet been tested for small waterbird detection from real-time surveillance videos, which can address the challenge of waterbird monitoring in real time. We propose an improved detection method by adding an extra prediction head, SimAM attention module, and sequential frame to YOLOv7, termed as YOLOv7-waterbird, for real-time video surveillance devices to identify attention regions and perform waterbird monitoring tasks. With the Waterbird Dataset, the mean average precision (mAP) value of YOLOv7-waterbird was 67.3%, which was approximately 5% higher than that of the baseline model. Furthermore, the improved method achieved a recall of 87.9% (precision = 85%) and 79.1% for small waterbirds (defined as pixels less than 40 × 40), suggesting a better performance for small object detection than the original method. This algorithm could be used by the administration of protected areas or other groups to monitor waterbirds with higher accuracy using existing surveillance cameras and can aid in wildlife conservation to some extent.

Using drone imagery to obtain population data of colony-nesting seabirds to support Canada’s transition to the global Key Biodiversity Areas program

Identifying of global or national biodiversity ‘hotspots’ has proven important for focusing and prioritizing conservation efforts worldwide. Canada has nearly 600 Important Bird and Biodiversity Areas (IBAs) identified by quantitative criteria to help guide avian conservation and management. Marine IBAs capture critical waterbird habitats such as nesting colonies, foraging sites, and staging areas. However, due to their remote locations, many lack recent population counts. Canada has begun transitioning IBAs into the global Key Biodiversity Areas (KBA) program; KBAs identify areas that are important for the persistence of biodiversity and encompass a wider scope of unique, rare, or vulnerable taxa. Assessing whether IBAs qualify as KBAs requires current data – as will future efforts to manage these biologically important sites. We conducted a pilot study in the Chain Islets and Great Chain Island IBA, in British Columbia, to assess the effectiveness of using drones to census surface-nesting seabirds in an IBA context. This IBA was originally designated for supporting a globally significant breeding colony of Glaucous-winged Gulls (Larus glaucescens). Total nest counts derived from orthomosaic imagery (1012 nesting pairs) show that this site now falls below the Global and National IBA designation criterion threshold, a finding consistent with regional declines in the species. Our trial successfully demonstrates a flexible and low cost approach to obtaining population data at an ecologically sensitive KBA site. We explore how drones will be a useful tool to assess and monitor species and habitats within remote, data-deficient IBAs, particularly during the transition to KBAs.

Artificial intelligence for automated detection of large mammals creates path to upscale drone surveys

Imagery from drones is becoming common in wildlife research and management, but processing data efficiently remains a challenge. We developed a methodology for training a convolutional neural network model on large-scale mosaic imagery to detect and count caribou (Rangifer tarandus), compare model performance with an experienced observer and a group of naïve observers, and discuss the use of aerial imagery and automated methods for large mammal surveys. Combining images taken at 75 m and 120 m above ground level, a faster region-based convolutional neural network (Faster-RCNN) model was trained in using annotated imagery with the labels: "adult caribou", "calf caribou", and "ghost caribou" (animals moving between images, producing blurring individuals during the photogrammetry processing). Accuracy, precision, and recall of the model were 80%, 90%, and 88%, respectively. Detections between the model and experienced observer were highly correlated (Pearson: 0.96-0.99, P value < 0.05). The model was generally more effective in detecting adults, calves, and ghosts than naïve observers at both altitudes. We also discuss the need to improve consistency of observers' annotations if manual review will be used to train models accurately. Generalization of automated methods for large mammal detections will be necessary for large-scale studies with diverse platforms, airspace restrictions, and sensor capabilities.

'Eye in the sky': Off-the-shelf unmanned aerial vehicle (UAV) highlights exposure of marine turtles to floating litter (FML) in nearshore waters of Mayo Bay, Philippines

Litter is a serious threat to the marine environment, with detrimental effects on wildlife and marine biodiversity. Limited data as a result of funding and logistical challenges in developing countries hamper our understanding of the problem. Here, we employed commercial unmanned aerial vehicle (UAV) as a cost-effective tool to study the exposure of marine turtles to floating marine litter (FML) in waters of Mayo Bay, Philippines. A quadcopter UAV was flown autonomously with on-board camera capturing videos during the flight. Still frames were extracted when either turtle or litter were detected in post-flight processing. The extracted frames were georeferenced and mapped using QGIS software. Results showed that turtles are highly exposed to FML in nearshore waters. Moreover, spatial dependence between FML and turtles was also observed. The study highlights the effectiveness of UAVs in marine litter research and underscores the threat of FML to turtles in nearshore waters.

A Review of Unoccupied Aerial Vehicle Use in Wetland Applications: Emerging Opportunities in Approach, Technology, and Data

Recent developments in technology and data processing for Unoccupied Aerial Vehicles (UAVs) have revolutionized the scope of ecosystem monitoring, providing novel pathways to fill the critical gap between limited-scope field surveys and limited-customization satellite and piloted aerial platforms. These advances are especially ground-breaking for supporting management, restoration, and conservation of landscapes with limited field access and vulnerable ecological systems, particularly wetlands. This study presents a scoping review of the current status and emerging opportunities in wetland UAV applications, with particular emphasis on ecosystem management goals and remaining research, technology, and data needs to even better support these goals in the future. Using 122 case studies from 29 countries, we discuss which wetland monitoring and management objectives are most served by this rapidly developing technology, and what workflows were employed to analyze these data. This review showcases many ways in which UAVs may help reduce or replace logistically demanding field surveys and can help improve the efficiency of UAV-based workflows to support longer-term monitoring in the face of wetland environmental challenges and management constraints. We also highlight several emerging trends in applications, technology, and data and offer insights into future needs.

Of Course We Fly Unmanned—We’re Women!

Striving to achieve a diverse and inclusive workplace has become a major goal for many organisations around the world [...]

Accuracy and precision of citizen scientist animal counts from drone imagery

Repeated counts of animal abundance can reveal changes in local ecosystem health and inform conservation strategies. Unmanned aircraft systems (UAS), also known as drones, are commonly used to photograph animals in remote locations; however, counting animals in images is a laborious task. Crowd-sourcing can reduce the time required to conduct these censuses considerably, but must first be validated against expert counts to measure sources of error. Our objectives were to assess the accuracy and precision of citizen science counts and make recommendations for future citizen science projects. We uploaded drone imagery from Año Nuevo Island (California, USA) to a curated Zooniverse website that instructed citizen scientists to count seals and sea lions. Across 212 days, over 1,500 volunteers counted animals in 90,000 photographs. We quantified the error associated with several descriptive statistics to extract a single citizen science count per photograph from the 15 repeat counts and then compared the resulting citizen science counts to expert counts. Although proportional error was relatively low (9% for sea lions and 5% for seals during the breeding seasons) and improved with repeat sampling, the 12+ volunteers required to reduce error was prohibitively slow, taking on average 6 weeks to estimate animals from a single drone flight covering 25 acres, despite strong public outreach efforts. The single best algorithm was ‘Median without the lowest two values’, demonstrating that citizen scientists tended to under-estimate the number of animals present. Citizen scientists accurately counted adult seals, but accuracy was lower when sea lions were present during the summer and could be confused for seals. We underscore the importance of validation efforts and careful project design for researchers hoping to combine citizen science with imagery from drones, occupied aircraft, and/or remote cameras.

Quantifying Waterfowl Numbers: Comparison of Drone and Ground-Based Survey Methods for Surveying Waterfowl on Artificial Waterbodies

Drones are becoming a common method for surveying wildlife as they offer an aerial perspective of the landscape. For waterbirds in particular, drones can overcome challenges associated with surveying locations not accessible on foot. With the rapid uptake of drone technology for bird surveys, there is a need to compare and calibrate new technologies with existing survey methods. We compared waterfowl counts derived from ground- and drone-based survey methods. We sought to determine if group size and waterbody size influenced the difference between counts of non-nesting waterfowl and if detection of species varied between survey methods. Surveys of waterfowl were carried out at constructed irrigation dams and wastewater treatment ponds throughout the Riverina region of New South Wales (NSW), Australia. Data were analyzed using Bayesian multilevel models (BMLM) with weakly informative priors. Overall, drone-derived counts of waterfowl were greater (+36%) than ground counts using a spotting scope (β_ground= 0.64 [0.62–0.66], (R2 = 0.973)). Ground counts also tended to underestimate the size of groups. Waterbody size had an effect on comparative counts, with ground counts being proportionally less than drone counts (mean = 0.74). The number of species identified in each waterbody type was similar regardless of survey method. Drone-derived counts are more accurate compared to traditional ground counts, but drones do have some drawbacks including initial equipment costs and time-consuming image or photo processing. Future surveys should consider using drones for more accurately surveying waterbirds, especially when large groups of birds are present on larger waterbodies.