Automated extraction of bank angles from bird‐borne video footage using open‐source software

The use of miniaturized video cameras to study the at‐sea behavior of flying seabirds has increased in recent years. These cameras allow researchers to record several behaviors that were not previously possible to observe. However, video recorders produce large amounts of data and videos can often be time‐consuming to analyze. We present a new technique using open‐source software to extract bank angles from bird‐borne video footage. Bank angle is a key facet of dynamic soaring, which allows albatrosses and petrels to efficiently search vast areas of ocean for food. Miniaturized video cameras were deployed on 28 Wandering Albatrosses (Diomedea exulans) on Marion Island (one of the two Prince Edward Islands) from 2016 to 2018. The OpenCV library for the Python programming language was used to extract the angle of the horizon relative to the bird’s body (= bank angle) from footage when the birds were flying using a series of steps focused on edge detection. The extracted angles were not significantly different from angles measured manually by three independent observers, thus being a valid method to measure bank angles. Image quality, high wind speeds, and sunlight all influenced the accuracy of angle estimates, but post‐processing eliminated most of these errors. Birds flew most often with cross‐winds (58%) and tailwinds (39%), resulting in skewed distributions of bank angles when birds turned into the wind more often. Higher wind speeds resulted in extreme bank angles (maximum observed was 94°). We present a novel method for measuring postural data from seabirds that can be used to describe the fine‐scale movements of the dynamic‐soaring cycle. Birds appeared to alter their bank angle in response to varying wind conditions to counter wind drift associated with the prevailing westerly winds in the Southern Ocean. These data, in combination with fine‐scale positional data, may lead to new insights into dynamic‐soaring flight.     

Testing the precision and sensitivity of density estimates obtained with a camera‐trap method revealed limitations and opportunities

The use of camera traps in ecology helps affordably address questions about the distribution and density of cryptic and mobile species. The random encounter model (REM) is a camera‐trap method that has been developed to estimate population densities using unmarked individuals. However, few studies have evaluated its reliability in the field, especially considering that this method relies on parameters obtained from collared animals (i.e., average speed, in km/h), which can be difficult to acquire at low cost and effort. Our objectives were to (1) assess the reliability of this camera‐trap method and (2) evaluate the influence of parameters coming from different populations on density estimates. We estimated a reference density of black bears (Ursus americanus) in Forillon National Park (Québec, Canada) using a spatial capture–recapture estimator based on hair‐snag stations. We calculated average speed using telemetry data acquired from four different bear populations located outside our study area and estimated densities using the REM. The reference density, determined with a Bayesian spatial capture–recapture model, was 2.87 individuals/10km² [95% CI: 2.41–3.45], which was slightly lower (although not significatively different) than the different densities estimated using REM (ranging from 4.06–5.38 bears/10km² depending on the average speed value used). Average speed values obtained from different populations had minor impacts on REM estimates when the difference in average speed between populations was low. Bias in speed values for slow‐moving species had more influence on REM density estimates than for fast‐moving species. We pointed out that a potential overestimation of density occurs when average speed is underestimated, that is, using GPS telemetry locations with large fix‐rate intervals. Our study suggests that REM could be an affordable alternative to conventional spatial capture–recapture, but highlights the need for further research to control for potential bias associated with speed values determined using GPS telemetry data.     

Three critical factors affecting automated image species recognition performance for camera traps

Ecological camera traps are increasingly used by wildlife biologists to unobtrusively monitor an ecosystems animal population. However, manual inspection of the images produced is expensive, laborious, and time‐consuming. The success of deep learning systems using camera trap images has been previously explored in preliminary stages. These studies, however, are lacking in their practicality. They are primarily focused on extremely large datasets, often millions of images, and there is little to no focus on performance when tasked with species identification in new locations not seen during training. Our goal was to test the capabilities of deep learning systems trained on camera trap images using modestly sized training data, compare performance when considering unseen background locations, and quantify the gradient of lower bound performance to provide a guideline of data requirements in correspondence to performance expectations. We use a dataset provided by Parks Canada containing 47,279 images collected from 36 unique geographic locations across multiple environments. Images represent 55 animal species and human activity with high‐class imbalance. We trained, tested, and compared the capabilities of six deep learning computer vision networks using transfer learning and image augmentation: DenseNet201, Inception‐ResNet‐V3, InceptionV3, NASNetMobile, MobileNetV2, and Xception. We compare overall performance on “trained” locations where DenseNet201 performed best with 95.6% top‐1 accuracy showing promise for deep learning methods for smaller scale research efforts. Using trained locations, classifications with <500 images had low and highly variable recall of 0.750 ± 0.329, while classifications with over 1,000 images had a high and stable recall of 0.971 ± 0.0137. Models tasked with classifying species from untrained locations were less accurate, with DenseNet201 performing best with 68.7% top‐1 accuracy. Finally, we provide an open repository where ecologists can insert their image data to train and test custom species detection models for their desired ecological domain.     

Factors influencing the diurnal behaviour of white rhino (Ceratotherium simum) in Matobo National Park,Zimbabwe

Activity periods and diurnal behaviour of southern white rhino (Ceratotherium simum simum) were studied in Matobo National Park, Zimbabwe, from August 2016 to July 2017 using camera trapping and focal group sampling, respectively. Camera trap data showed that diurnal and nocturnal movements of white rhino varied significantly with moon phase, with more frequent nocturnal photographic captures during full- and third-quarter moon phases than during new and crescent moon phases. Permutational MANOVA and canonical correspondence analysis of focal group observation data were used to test the effects of season, time of day, habitat type, distance to roads, moon phase, presence/absence of other mammals, age/sex category, individual identity and herd composition on daytime behaviour. Weather conditions accounted for the majority of variation in rhino behaviour, with more time spent feeding during cloudy and rainy weather than sunny weather. Time of day and moon phase also accounted for high variation in behaviour, followed by habitat attributes and age/sex category. Variance partitioning by sets of explanatory variables showed that weather conditions had the greatest explanatory power in the variation of rhino behaviour, followed by temporal attributes, habitat attributes and age/sex category. Our findings provide insights for white rhino population management strategies aimed at increasing the effectiveness of security patrols, minimising disturbance and improving sightings for tourists.     

Human disturbance and the activity patterns and temporal overlap of tapirs and jaguars in reserves of NW Belize

Human disturbance from tourism and other non-consumptive activities in protected areas may be stressful to wildlife. Animals may move away in space or time to avoid human interaction. For species of particular conservation concern, such as Baird's tapirs (Tapirus bairdii) and jaguars (Panthera onca), a better understanding of how they respond to different levels and types of disturbance is needed in order to manage human visitation to parks in ways that minimize negative outcomes for wildlife. We describe the overlap in activity patterns of tapirs, jaguars, and humans at logged and unlogged sites and at places with low versus high human visitation using camera survey data from protected areas of NW Belize, 2013–2016. Tapirs were nocturnal in all study sites, with > 80% of all tapir detections occurring between 1900 hr and 0500 hr. Their activity patterns were not different in unlogged versus logged sites and did not change with increased human traffic. Jaguars were cathemeral across sites but had more nocturnal activity at the site with the most human impact. Activity pattern overlap between tapirs and jaguars did not differ significantly between logged and unlogged sites, nor between areas with low and high human activity. Human traffic increased from 2013 to 2016 at most of the study locations. In conclusion, this camera trap dataset suggests that non-consumptive human disturbance does not alter the activity patterns of tapirs and jaguars in protected areas lacking hunting pressure.     

利用红外线相机监测地表水对广西弄岗国家级自然保护区兽类分布的影响

广西弄岗国家级自然保护区属岩溶地貌,表土层薄,储水度低,地表水奇缺,动物可能因水源而改变其活动范围和规律。2006年3月—2007年6月和2010年10月—2011年5月,该实验利用红外线相机在兽类通往水源地之间,即兽径上(如足迹和粪便)进行兽类调查。共拍摄到兽类5目12科17属19种,包括国家一级保护动物黑叶猴(Trachypithecus francoisi)和熊猴(Macaca assamensis)。物种累积曲线95%置信区间重迭,相似性分析和非度量多维标度显示两种微生境出没物种的差异较小,然而,差异物种占整体记录的42%,可见岩溶地区水资源极其珍贵。中、短期保护建议包括加强水源点的巡护、防止偷猎、限制滥抽地下水,以及考虑整合呈岛屿状的陇呼、弄岗及陇山地区,以加强保护大面积生境的连贯性,并加速恢复山峰间已遭破坏的圆洼谷地的原生树林。     

Changing use of camera traps in mammalian field research: habitats,taxa and study types

相似文献   

Does underwater flash photography affect the behaviour,movement and site persistence of seahorses?

The effect of flash photography on seahorse species has never been tested. An experiment was established to test the effect of flash photography and the handling of Hippocampus whitei, a medium‐sized seahorse species endemic to Australia, on their behavioural responses, movements and site persistence. A total of 24 H. whitei were utilized in the experiment with eight in each of the three treatments (flash photography, handling and control). The effect of underwater flash photography on H. whitei movements was not significant; however, the effect of handling H. whitei to take a photograph had a significant effect on their short‐term behavioural responses to the photographer. Kaplan–Meier log‐rank test revealed that there was no significant difference in site persistence of H. whitei from each of the three treatments and that flash photography had no long‐term effects on their site persistence. It is concluded that the use of flash photography by divers is a safe and viable technique with H. whitei, particularly if photographs can be used for individual identification purposes.     

高山兀鹫繁殖期交配及筑巢行为初步观察

高山兀鹫(Gyps himalayensis)在海拔2 400~4 800 m的悬崖峭壁之上营巢,行为观测难度较大,资料匮缺。2013年至2015年,在天山采用样线调查法、问卷调查法,寻找到上百个巢穴。通过布设红外自动相机的方法,对高山兀鹫繁殖期巢内行为进行拍摄,同时结合焦点动物取样法和瞬时扫描法,对其交配及巢内行为进行观察和研究。共拍摄11万张照片,选出有效照片2 150张。对繁殖期行为进行了描述和定义,共7大类16种行为。根据照片逐一对比每种行为,统计各种行为拍摄张数及所占比例,其中,交配135张(6.3%),筑巢403张(18.7%),休息420张(19.5%),警戒375张(17.4%),保养335张(15.6%),运动200张(9.3%),其他282张(13.1%)。照片记录显示,交配行为从1月16日持续到3月2日,在一天的8:00~16:00时间段内,除9:00~10:00时没有记录到交配行为外,其余时间段内都有发生。交配行为次数(张数)在13:00~14:00时达到高峰,与气温同期达到高峰。交配过程持续8~15 s。筑巢期间,巢材由雌雄兀鹫运输,雌性负责将巢材放入巢中,整理、铺垫和修缮。筑巢行为在白天有三个高峰期,分别为11:00~12:00时、13:00~14:00时和15:00~16:00时。在每个高峰期后,往往出现一个短暂的低谷期。冬季食物短缺,照片显示雕鸮(Bubo bubo)、雪豹(Uncia uncia)等食肉动物进入巢内,对兀鹫繁殖造成干扰,甚至中断繁殖。作为高原健康指示物种,高山兀鹫生存状况不容乐观,急需加强保护。     

贵州赤水桫椤国家级自然保护区鸟兽多样性红外相机初步监测

红外相机监测是了解野生动物多样性和威胁因素的重要手段。本研究采用网格法和分层抽样调查法, 在贵州赤水桫椤国家级自然保护区内选取20个监测位点布设红外相机, 对区内鸟兽物种多样性进行监测。2015年8月至2017年8月, 红外相机累计工作6,370个工作日, 共拍摄45,953张照片, 独立有效照片1,936张。准确鉴定出兽类4目8科19种, 鸟类4目11科28种, 其中, 国家II级重点保护野生动物7种。相对丰富度指数(RAI)排前五位的兽类依次是毛冠鹿(Elaphodus cephalophus)、鼬獾(Melogale moschata)、藏酋猴(Macaca thibetana)、小麂(Muntiacus reevesi)和野猪(Sus scrofa); 鸟类依次是紫啸鸫(Myophonus caeruleus)、红腹角雉(Tragopan temminckii)、灰胸竹鸡(Bambusicola thoracicus)、黑喉噪鹛(Garrulax chinensis)和棕颈钩嘴鹛(Pomatorhinus ruficollis)。物种积累曲线结果表明, 兽类稀疏化曲线在300天后趋于稳定, 表明监测取样已较充分, 而鸟类监测物种数随时间积累依旧保持增长趋势。