科学家们首次利用卫星图像统计非洲大象的数量。照片由一颗在地表 600 公里（ 372 英里）外轨道上运行的地球观测卫星拍摄。这一突破可以让人们在无云日里对 5000 平方公里的大象栖息地进行观测。
When you first look at it, the image is a series of grey blobs amid a forest of green splotches. But when you study it closely, the picture the researchers are using, taken from space, is of a herd1 of elephants wandering through the trees.
乍眼一看，这些图像是绿色斑点状森林中的一块块灰色团状物。但当你仔细观察后会发现，研究人员们所使用的这张从太空中拍摄的照片其实是一群在林间漫步的大象。
The scientists used a series of these images – pictures of one National Park in South Africa, taken by a satellite – to test a new way of monitoring the African elephant population.
科学家们利用了一系列由卫星拍摄的南非一个国家公园的图像，来测试一种用于监测非洲大象数量的新方法。
The laborious2 task of actually counting the animals was all done via machine learning. By essentially3 showing a computer dozens of these pictures, the researchers trained it to correctly pick out and count elephants amid trees, rocks and other animals.
实际统计动物数量的艰巨任务是通过机器学习算法完成的。实际上就是通过向电脑展示几十张这样的照片，研究人员训练计算机在树木、岩石和其它动物中正确识别并清点大象的数量。
网友评论