Aerial photography and machine learning for estimating extremely high flamingo numbers on the Makgadikgadi Pans, Botswana

Sophie Yang, Roxane J. Francis, Mike Holding, Richard T. Kingsford, Aerial photography and machine learning for estimating extremely high flamingo numbers on the Makgadikgadi Pans, Botswana, Global Ecology and Conservation, 2024, e03011, ISSN 2351-9894, https://doi.org/10.1016/j.gecco.2024.e03011. (https://www.sciencedirect.com/science/article/pii/S2351989424002154) We developed a semi-supervised machine learning method for counting a large feeding concentration of flamingos (2 June 2019) in aerial photographs from northern Sua Pan of the Makgadikgadi Pans. We also analysed rainfall and flooding frequency and extent, using satellite imagery, estimating likely frequency of these flamingo concentrations. Our analysis successfully provided an estimate of 372,172 to 689,473 flamingos, with methods producing over 97% test accuracy. Uncertainty related primarily to data coverage and collection rather than methodology. Code from Google Earth Engine has been provided in a JavaScript file, along with an example collage. To use the code, please copy and paste into Google Earth Engine scripts and upload relevant images as assets to be analysed.

杨索菲、罗克桑·J·弗朗西斯、迈克·霍尔丁、理查德·T·金斯福德. 航空摄影与机器学习用于估算博茨瓦纳马卡迪卡迪盐沼的极大量火烈鸟数量[J]. 《全球生态学与保护（Global Ecology and Conservation）》, 2024, e03011. ISSN 2351-9894. https://doi.org/10.1016/j.gecco.2024.e03011. (https://www.sciencedirect.com/science/article/pii/S2351989424002154) 我们开发了一种半监督机器学习方法，用于计数博茨瓦纳马卡迪卡迪盐沼北部苏阿潘地区2019年6月2日拍摄的航空照片中的大型火烈鸟觅食集群。我们还利用卫星影像分析了降雨、洪水发生频率与影响范围，以此估算此类火烈鸟集群的可能出现频率。本研究的分析结果成功给出了372172至689473只火烈鸟的数量估算值，相关方法的测试准确率超过97%。其不确定性主要源于数据覆盖范围与采集环节，而非方法本身。我们已将谷歌地球引擎（Google Earth Engine）的代码以JavaScript文件形式提供，并附带示例拼接图。使用该代码时，请将代码复制粘贴至谷歌地球引擎脚本中，并将待分析的相关影像作为资产上传。