当前位置: X-MOL 学术Remote Sens. Ecol. Conserv. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Machine learning to detect marine animals in UAV imagery: effect of morphology, spacing, behaviour and habitat
Remote Sensing in Ecology and Conservation ( IF 3.9 ) Pub Date : 2021-05-05 , DOI: 10.1002/rse2.205
Antoine M. Dujon 1 , Daniel Ierodiaconou 2 , Johanna J. Geeson 3 , John P. Y. Arnould 3 , Blake M. Allan 2 , Kostas A. Katselidis 4 , Gail Schofield 1, 5
Affiliation  

Machine learning algorithms are being increasingly used to process large volumes of wildlife imagery data from unmanned aerial vehicles (UAVs); however, suitable algorithms to monitor multiple species are required to enhance efficiency. Here, we developed a machine learning algorithm using a low-cost computer. We trained a convolutional neural network and tested its performance in: (1) distinguishing focal organisms of three marine taxa (Australian fur seals, loggerhead sea turtles and Australasian gannets; body size ranges: 0.8–2.5 m, 0.6–1.0 m, and 0.8–0.9 m, respectively); and (2) simultaneously delineating the fine-scale movement trajectories of multiple sea turtles at a fish cleaning station. For all species, the algorithm performed best at detecting individuals of similar body length, displaying consistent behaviour or occupying uniform habitat (proportion of individuals detected, or recall of 0.94, 0.79 and 0.75 for gannets, seals and turtles, respectively). For gannets, performance was impacted by spacing (huddling pairs with offspring) and behaviour (resting vs. flying shapes, overall precision: 0.74). For seals, accuracy was impacted by morphology (sexual dimorphism and pups), spacing (huddling and creches) and habitat complexity (seal sized boulders) (overall precision: 0.27). For sea turtles, performance was impacted by habitat complexity, position in water column, spacing, behaviour (interacting individuals) and turbidity (overall precision: 0.24); body size variation had no impact. For sea turtle trajectories, locations were estimated with a relative positioning error of <50 cm. In conclusion, we demonstrate that, while the same machine learning algorithm can be used to survey multiple species, no single algorithm captures all components optimally within a given site. We recommend that, rather than attempting to fully automate detection of UAV imagery data, semi-automation is implemented (i.e. part automated and part manual, as commonly practised for photo-identification). Approaches to enhance the efficiency of manual detection are required in parallel to the development of effective implementation of machine learning algorithms.

中文翻译:

机器学习在无人机图像中检测海洋动物:形态、间距、行为和栖息地的影响

机器学习算法越来越多地用于处理来自无人机 (UAV) 的大量野生动物图像数据;然而,需要合适的算法来监测多个物种以提高效率。在这里,我们使用低成本计算机开发了机器学习算法。我们训练了一个卷积神经网络并测试了其在以下方面的性能:(1)区分三种海洋分类群(澳大利亚海豹、红海龟和澳大利亚塘鹅;体型范围:0.8-2.5 m、0.6-1.0 m 和 0.8 –0.9 m,分别);(2)同时在鱼类清洗站勾勒出多只海龟的精细运动轨迹。对于所有物种,该算法在检测体长相似的个体方面表现最佳,表现出一致的行为或占据统一的栖息地(检测到的个体比例,或者塘鹅、海豹和海龟的召回率分别为 0.94、0.79 和 0.75)。对于塘鹅,性能受到间距(与后代挤在一起)和行为(静止与飞行形状,整体精度:0.74)的影响。对于海豹,准确性受到形态(性二态性和幼崽)、间距(蜷缩和托儿所)和栖息地复杂性(海豹大小的巨石)(总体精度:0.27)的影响。对于海龟,性能受到栖息地复杂性、水柱位置、间距、行为(相互作用的个体)和浊度(总体精度:0.24)的影响;体型变化没有影响。对于海龟轨迹,估计位置的相对定位误差 <50 厘米。总之,我们证明,虽然相同的机器学习算法可用于调查多个物种,但没有一种算法可以最佳地捕获给定地点内的所有组件。我们建议,与其尝试完全自动化无人机图像数据的检测,不如实施半自动化(即部分自动和部分手动,如通常用于照片识别的做法)。在开发有效实施机器学习算法的同时,还需要提高手动检测效率的方法。如通常用于照片识别的做法)。在开发有效实施机器学习算法的同时,还需要提高手动检测效率的方法。如通常用于照片识别的做法)。在开发有效实施机器学习算法的同时,还需要提高手动检测效率的方法。
更新日期:2021-05-05
down
wechat
bug