US 10,872,417 B1Granted

Automatic delineation agricultural field management zones using remote sensing and field data

AssigneeFlyPard Analytics GmbH

InventorsDzmitry Yablonski, Dmitry Dementiev, Vladimir Klinkov

Life of this patent

Estimating the term…

Loading patent…

Abstract

A method for automatic delineation of agricultural field management zones using remote sensing data from multiple years of observation and field data, including: selecting and processing remote sensing images appropriate for management zones delineation; defining management zones analytics parameters; generating a dataset by calculating vegetation indices for every appropriate remote sensing image, by removing outliers and values of non-arable spots, by considering different resolutions of remote sensing images, by applying statistical interpolation algorithms; generating field management zones by applying statistical classification methods on the dataset including merging small and separate management zones into larger neighboring zones. A computer system is programmed and configured to generate field management zones automatically; to enable user to assign rates for every management zone, and to export them to agricultural machinery and to mobile phone or tablet device for online and offline usage.

Claims14

Dependent claims

A method for automatic delineation of agricultural field management zones, comprising the steps of:

selecting the remote sensing images that satisfy the following criteria (i) no clouds or cloud shadows are present, (ii) autodetection of artificial in-field objects that can be processed out of the images, (iii) average vegetation index per crop per region is in a specified range, and (iv) images were taken during a peak vegetation period for a particular crop in that region;
calculating an optimal number of zones in the agricultural field based on a size of the agricultural field area and in-field vegetation variability between most productive and least productive spots;
calculating an optimal resolution pixel size of an output map of the field management zones based on resolution of the selected remote sensing images;
reducing size of pixel extraction field area in selected remote sensing images by omitting pixels corresponding edges of the agricultural field area, and taking into account the resolution of the images;
and
generating the field management zones, by
(i) overlaying a grid on the selected remote sensing images;
(ii) calculating vegetation index at each point of the grid;
(iii) at each point of the grid, removing outliers and values of non-arable spots to generate an analytic dataset for each selected image;
(iv) merging analytic datasets of images that have the same resolution;
(v) increasing a resolution of the merged analytic datasets including filling of any gaps in a surface of the field management zones;
(vi) merging analytic datasets corresponding to different providers into a single dataset;
and
(vii) generating the output map of the field management zones by splitting pixels of the single dataset into groups based on their values, and merging separated polygons into larger polygons, and then cropping the field management zones based on original field boundaries.

11 more elements

A data processing system comprising:

one or more servers configured to perform the following steps:
selecting the remote sensing images that satisfy the following criteria (i) no clouds or cloud shadows are present, (ii) autodetection of artificial in-field objects that can be processed out of the images, (iii) average vegetation index per crop per region is in a specified range, and (iv) images were taken during a peak vegetation period for a particular crop in that region;
calculating an optimal number of zones in the agricultural field based on a size of the agricultural field area and in-field vegetation variability between most productive and least productive spots;
calculating an optimal resolution pixel size of an output map of the field management zones based on resolution of the selected remote sensing images;
reducing size of the pixel extraction field area in remote sensing images by omitting pixels corresponding edges of the agricultural field area, and taking into account the resolution of the images;
and
generating the field management zones, by
(i) overlaying a grid on the selected remote sensing images;
(ii) calculating vegetation index at each point of the grid;
(iii) at each point of the grid, removing outliers and values of non-arable spots to generate an analytic dataset for each selected image;
(iv) merging analytic datasets of images that have the same resolution;
(v) increasing a resolution of the merged analytic datasets, including filling of any gaps in a surface of the field management zones;
(vi) merging analytic datasets corresponding to different providers into a single dataset;
and
(vii) generating the output map of the field management zones by splitting pixels of the single dataset into groups based on their values, and merging separated polygons into larger polygons, and then cropping the field management zones based on original field boundaries.

12 more elements

Dependencies are inferred from the claim text.