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.)
21 000 birds in 4.5 h: efficient large-scale seabird detection with machine learning
Remote Sensing in Ecology and Conservation ( IF 3.9 ) Pub Date : 2021-03-25 , DOI: 10.1002/rse2.200
Benjamin Kellenberger 1, 2 , Thor Veen 3, 4 , Eelke Folmer 4 , Devis Tuia 2
Affiliation  

We address the task of automatically detecting and counting seabirds in unmanned aerial vehicle (UAV) imagery using deep convolutional neural networks (CNNs). Our study area, the coast of West Africa, harbours significant breeding colonies of terns and gulls, which as top predators in the food web function as important bioindicators for the health of the marine ecosystem. Surveys to estimate breeding numbers have hitherto been carried out on foot, which is tedious, imprecise and causes disturbance. By using UAVs and CNNs that allow localizing tens of thousands of birds automatically, we show that all three limitations can be addressed elegantly. As we employ a lightweight CNN architecture and incorporate prior knowledge about the spatial distribution of birds within the colonies, we were able to reduce the number of bird annotations required for CNN training to just 200 examples per class. Our model obtains good accuracy for the most abundant species of royal terns (90% precision at 90% recall), but is less accurate for the rarer Caspian terns and gull species (60% precision at 68% recall, respectively 20% precision at 88% recall), which amounts to around 7% of all individuals present. In sum, our results show that we can detect and classify the majority of 21 000 birds in just 4.5 h, start to finish, as opposed to about 3 weeks of tediously identifying and labelling all birds by hand.

中文翻译:

4.5 小时内检测 21 000 只鸟:使用机器学习进行高效的大规模海鸟检测

我们使用深度卷积神经网络 (CNN) 解决了在无人驾驶飞行器 (UAV) 图像中自动检测和计数海鸟的任务。我们的研究区,西非海岸,拥有大量燕鸥和海鸥的繁殖群,它们作为食物网中的顶级捕食者,是海洋生态系统健康的重要生物指标。迄今为止,估计繁殖数量的调查都是步行进行的,这既乏味又不精确,而且会造成干扰。通过使用允许自动定位数万只鸟类的无人机和 CNN,我们表明可以优雅地解决所有三个限制。由于我们采用了轻量级的 CNN 架构并结合了关于群体内鸟类空间分布的先验知识,我们能够将 CNN 训练所需的鸟类注释数量减少到每类仅 200 个示例。我们的模型对最丰富的皇家燕鸥种类获得了良好的准确度(在 90% 的召回率下达到了 90% 的准确率),但对于稀有的里海燕鸥和海鸥种类的准确度较低(在 68% 的召回率下准确率为 60%,在 88 时准确率分别为 20% % 回忆),约占所有在场人员的 7%。总而言之,我们的结果表明,我们可以在 4.5 小时内(从头到尾)检测和分类 21 000 只鸟类中的大部分,而手动识别和标记所有鸟类需要大约 3 周的繁琐时间。但对于更稀有的里海燕鸥和海鸥物种(68% 召回率为 60%,召回率为 20%,召回率为 20%),这相当于所有在场个体的 7% 左右。总而言之,我们的结果表明,我们可以在 4.5 小时内(从头到尾)检测和分类 21 000 只鸟类中的大部分,而手动识别和标记所有鸟类需要大约 3 周的繁琐时间。但对于更稀有的里海燕鸥和海鸥物种(68% 召回率为 60%,召回率为 20%,召回率为 20%),这相当于所有在场个体的 7% 左右。总而言之,我们的结果表明,我们可以在 4.5 小时内(从头到尾)检测和分类 21 000 只鸟类中的大部分,而手动识别和标记所有鸟类需要大约 3 周的繁琐时间。
更新日期:2021-03-25
down
wechat
bug