Vegetation indices have become an indispensable tool in remote sensing‐based agricultural research. A recent area of advancement in agricultural remote sensing research is in high‐throughput phenotyping, often conducted on a plot by plot basis. FIELDimageR is a tool used extensively in high‐throughput phenotyping that estimates zonal statistics of vegetation indices per plot. However, being written in R language, FIELDimageR requires high computing time. As a high‐resolution image over a large area means a large number of pixels, FIELDimageR is incapable of using high‐resolution orthomosaicked images without reducing image resolution by aggregating digital numbers of several pixels and treating them as one pixel. This research tool implements FIELDimageR in the Python language as FIELDimagePy. FIELDimagePy follows similar workflows as FIELDimageR and generates equivalent results for zonal statistics of vegetation indices per plot. FIELDimagePy is significantly and substantially faster than FIELDimageR. Computing time by FIELDimagePy are three to four times lower than computing times by FIELDimageR, even when using raw images with 16 times denser pixels. Moreover, FIELDimagePy is useful beyond plot by plot research in agriculture and capable of estimating zonal statistics of any raster bounded by any polygons. With slight modifications, FIELDimagePy can be useful for other disciplines of science, such as geophysics, geography, economics, medical sciences, among others. FIELDimagePy can be accessed from the GitHub repository: https://github.com/SumantraChatterjee/FIELDimagePy.

中文翻译：

---

FIELDimagePy：一种用于估计图像区域统计数据的工具，以一个或多个多边形为界


植被指数已成为基于遥感的农业研究中不可或缺的工具。农业遥感研究的最新进展领域是高通量表型分析，通常逐个地块进行。FIELDimageR 是一种广泛用于高通量表型分析的工具，可估计每个样地植被指数的分区统计。但是，由于是用 R 语言编写的，FIELDimageR 需要较长的计算时间。由于大面积上的高分辨率图像意味着大量像素，因此 FIELDimageR 无法通过聚合多个像素的数字数字并将其视为一个像素来降低图像分辨率。此研究工具在 Python 语言中将 FIELDimageR 实现为 FIELDimagePy。FIELDimagePy 遵循与 FIELDimageR 类似的工作流程，并为每个图的植被指数分区统计生成等效结果。FIELDimagePy 比 FIELDimageR 快得多。FIELDimagePy 的计算时间比 FIELDimageR 的计算时间短 3 到 4 倍，即使使用像素密度高 16 倍的原始图像也是如此。此外，FIELDimagePy 在农业中不仅仅逐个地块研究，并且能够估计由任何多边形界定的任何栅格的分区统计数据。稍作修改，FIELDimagePy 可用于其他科学学科，例如地球物理学、地理学、经济学、医学科学等。FIELDimagePy 可以从 GitHub 存储库访问：https://github.com/SumantraChatterjee/FIELDimagePy。