Method for correcting the time delay in measuring agricultural yield

• US 9,972,058 B2
• Filed: 03/20/2017
• Issued: 05/15/2018
• Est. Priority Date: 06/29/2011
• Status: Active Grant

• Current Assignee: GEOSYS SA (UrtheCast Corp.)
• Original Assignee: GEOSYS SA (UrtheCast Corp.)
• Inventors: Romier, Chloe
• Primary Examiner(s): Yi, Roy Y
• Application Number:
  US15/463,442
• Publication Number:
  US 20170193613A1
• Time in Patent Office: 421 Days
• CPC Class Codes: A01B 79/005 Precision agriculture
• CPC Class Codes: A01D 41/1271 for measuring crop flow
• CPC Class Codes: G01N 33/0098 Plants or trees wood G01N33/46
• CPC Class Codes: G06Q 50/02 Agriculture; Fishing; Forestry; Mining

Abstract

The invention relates to a method of processing a set of data collected by at least one sensor at successive points on the travel over a plot of land (1), by at least one vehicle (2), each one of the said pieces of data having been recorded with a time difference in relation to the moment of the vehicle passing over a given point on the plot, i.e. representing information relating to the state of the vehicle at a moment prior to its moment of recording, each piece of data being associated with a moment of recording and a geographical position of a point known as the measuring point at this moment of recording, supplied by geolocation means.

The method comprises a phase of estimating an optimum difference value by minimizing a degree of projection of a three-dimensional map of values measured at each geographical point, and correction of the data measured from this difference.

Claims

1. A method of correcting a time lag for a set of data, comprising the steps of:

generating a plurality of adjusted data sets by replacing geographic coordinates of a measuring point supplied by a geolocation system with geographic coordinates of a previous or a future measuring point having a time lag value to shift the geographic coordinates of the measuring point; and

for each set of adjusted data;

building a three-dimensional map showing latitudes, longitudes and heights along the X, Y and Z axes, respectively, in accordance with the adjusted data associated with the measuring point, wherein a height represents a sensor value;

determining a measurement of a developed surface area of the three-dimensional map;

calculating a surface area of a projection of the three-dimensional map onto a plane of latitudes and longitudes;

calculating a projection ratio between the developed surface area of the three-dimensional map and the area of projection;

determining a minimum value of the projection ratio for the sets of adjusted data; and

correcting the set of data using the minimum value.

2. The method of claim 1, further comprising the steps of selecting, within each set of adjusted data, a set of normal measuring points comprising the measuring points positioned on substantially straight rows of lines parallel to one another;

and excluding measuring points located at a start and an end of each of the rows, located at half-turns and constituting an identification of a plot of land from the set of normal measuring points.

3. The method of claim 2, further comprising the steps of projecting the normal measuring points onto a regular mesh grid;

and interpolating associated data by linear regression.

4. The method claim 1, further comprising the step of filtering the corrected data using the minimum value and wherein the step of filtering further comprises the steps of:

eliminating points in the corrected data having a weighted flow or yield value equal to 0;

calculating an average M and standard deviation EC of flow measurements or yield values;

eliminating measuring points having an absolute flow or yield value higher than the absolute value of the average M plus k times the standard deviation EC for the flow or yield data, respectively.

5. The method of claim 1, wherein the data are crop flow or yield measurements;

and further comprising the step of weighing the crop flow or yield measurements from a total weight harvested from a plot of land.

6. The method of claim 1, wherein the step of generating the set of adjusted data to shift the geographic coordinates of the measuring point comprises the step of shifting values recorded by the sensor from one measuring point to another.

7. A method, comprising:

receiving a data set comprising a set of geographic coordinates defining a plurality of measuring points and sensor data collected at each of the measuring points by a vehicle moving over a plot of land, the measuring points and the sensor data being misaligned due to a time lag of unknown duration in the sensor data;

generating an adjusted data set by correcting the data set based on an optimal value for the time lag, the optimal value for the time lag being one of a plurality of candidate time values and having been determined to minimize the sensor data misalignment as compared to other candidate time lag values of the plurality of candidate time values; and

generating a yield map by distributing sensor data associated with each particular measuring point of the adjusted data set across an area associated with the particular measuring point and adding the distributed sensor data to a grid of cells that is applied over a representation of the plot of land.

8. The method of claim 7, further comprising:

importing a set of raw data files generated from a complete harvesting of the plot of land;

translating the set of raw data files into a set of formatted data files in a standard format;

processing the set of formatted data files to generate a set of plot files each including information connected to a single farming plot, a single crop, and a single crop cycle.

9. The method of claim 7, wherein:

the geographic coordinates of the measuring points are determined by a global positioning system as the vehicle moves over the plot of land; and

the sensor data associated with a particular measuring point is determined by at least one sensor as the sensor passes over a geographical point associated with the geographic coordinates associated with the measuring point, the at least one sensor being carried by the vehicle as the vehicle moves over the plot of land.

10. The method of claim 7, wherein the sensor data includes flow values, the method further comprising weighting the flow values measured at each measuring point in accordance with a total crop weight harvested on the plot of land.

11. The method of claim 7, wherein determining the optimal value of the time lag comprises:

generating an adjusted data set for each candidate time value by replacing the geographic coordinates of each measuring point in the data set with geographic coordinates of previous or future measuring points based on the candidate time value to shift the measuring points;

determining an amount of data misalignment for each adjusted data set; and

selecting as the optimal value for the time lag a candidate time value corresponding to an adjusted data set that minimizes the amount of data misalignment.

12. The method of claim 11, wherein determining an amount of data misalignment comprises:

selecting a set of normal measuring points from the measuring points of the adjusted data set;

projecting the set of normal measuring points onto a regular mesh grid to obtain a set of meshed points;

determining a three dimensional meshed data map based on the set of meshed points;

calculating a developed area as a sum of an area of sides of a polygon defined by the three dimensional meshed data map;

calculating a projected area corresponding to an area of a projection of the three dimensional meshed data map onto projection plan defined by three dimensional meshed data map;

calculating a projection rate as a ratio between the developed area and the projected area; and

defining the projection rate as a measure of the amount of data misalignment in the adjusted data set.

13. The method of claim 12, wherein the step of selecting normal measuring points includes selecting measuring points positioned on substantially straight rows of parallel lines that correspond to opposing directions of travel of the vehicle over the plot of land.

14. The method of claim 12, wherein the three dimensional meshed data map includes X,Y coordinates that define geographic locations and Z coordinates that define weighted flow values associated with each geographic coordinate.

15. The method of claim 12, wherein the three dimensional meshed data map includes X,Y coordinates that define geographic locations and Z coordinates that define weighted yield values associated with each geographic coordinate.

16. The method of claim 7, wherein a given measuring point is associated with sensor data collected across of a width of a cutting bar that is coupled to the vehicle, the method further comprising distributing sensor data over a space covered by the width of the cutting bar.

17. The method of claim 16, wherein distributing sensor data includes replacing a given measuring point with a plurality of points distributed homogenously over the width of the cutting bar.

18. The method of claim 16, wherein distributing sensor data includes replacing a given measuring point with a plurality of points distributed homogenously between two consecutive moments of measurement.

19. The method of claim 7, further comprising:

applying a grid of predetermined mesh over the representation of the plot of land;

adding together flows measured at all points inside each cell of the grid;

associating an output measured at a given frequency with each point inside each cell of the grid to define a measured flow; and

dividing the measured flow by the area of the cell to obtain a yield value associated with each cell.

20. The method of claim 19, wherein the representation of the plot of land is a satellite image.

21. The method of claim 19, further comprising:

defining a percentage of area occupied by the plot of land within an edge cell by intersecting a harvested plot contour with cell edges.

22. The method of claim 21, further comprising:

creating a set of buffers by cutting a half-width around each measuring point; and

merging together and smoothing the set of buffers to obtain the harvested plot contour.