Methods and systems for use in processing images related to crop phenology

1 . A computer-implemented method for use in processing image data associated with one or more plots, the method comprising:

accessing, by a computing device, a data set, the data set including images associated with one or more plots, the images having a spatial resolution of better than or equal to one foot by one foot per pixel;

for each of the one or more plots:

comparing, by the computing device, for each pixel in the images, a first normalized difference vegetation index (NDVI) value of a first image of the plot at time n from the data set to a second NDVI value of a second image of the plot at time n+1 from the data set;

determining, by the computing device, differences of the second NDVI values relative to the first NDVI values;

in response to a first threshold percentage of the differences being positive, flagging, by the computing device, at least part of the second image as bad data; and

modifying, by the computing device, the data set by removing the at least part of the second image, based on the second image being flagged as bad data, from the data set;

preparing the modified data set as a series of feature vectors having a length of at least m*n, where m is a number of spectral values included in each of said images in the modified data set, and n is a number of image dates which includes said images for the one or more plots in the modified data set, the spectral values including spectral color data from said images and NDVI values, and wherein the feature vectors further include at least one of the differences of the second NDVI values relative to the first NDVI values;

accessing phenotypic data for a crop included in the one or more plots, at a time consistent with the images associated with the one or more plots, the phenotypic data including relative maturity;

training a model, based on data including the series of feature vectors for the one or more plots, and the accessed phenotypic data, to identify relative maturity of a target crop in a target plot from images of the target plot, whereby weights of the model are fitted to said data; and

storing, by the computing device, the trained model in memory.

2 . The computer-implemented method of claim 1 , wherein comparing the first NDVI value of the first image of the plot at time n to the second NDVI value of the second image of the plot at time n+1 includes comparing each NDVI value of the first image to corresponding NDVI values of the second image, where the NDVI values correspond based on a corresponding location in the plot; and

wherein flagging the second image in response to the second NDVI value being greater than the first NDVI value includes flagging the second image in response to each NDVI value of the second image being greater than each corresponding NDVI value of the first image.

3 . The computer-implemented method of claim 1 , wherein the model is one of: a random forest algorithm and a boosting algorithm.

4 . The computer-implemented method of claim 1 , wherein the series of feature vectors further includes a NDVI delta, wherein the NDVI delta includes a difference between the NDVI values of consecutive images in the data set.

5 . The computer-implemented method of claim 1 , further comprising forecasting phenotypic data for a plot based on the trained model.

6 . The computer-implemented method of claim 1 , further comprising generating, via the model, phenotypic data for the crop planted in at least one of the one or more plots.

7 . The computer-implemented method of claim 6 , further comprising:

calculating, by the computing device, a confidence interval for the model based on the generated phenotypic data for the crop planted in the at least one of the one or more plots; and

storing, by the computing device, the calculated confidence interval in the memory in association with the model.

8 . A system for use in processing image data associated with one or more plots, the system comprising a computing device including a processor, which is configured, by executable instructions, to:

access a data set, the data set including images associated with one or more plots, the images having a spatial resolution of better than or equal to one foot by one foot per pixel;

for each of the one or more plots:

compare, for each pixel in the images, a first normalized difference vegetation index (NDVI) value of a first image of the plot at time n from the data set to a second NDVI value of a second image of the plot at time n+1 from the data set;

determine differences of the second NDVI values relative to the first NDVI values;

in response to a first threshold percentage of the differences being positive, flag at least part of the second image; and

modify the data set by removing the at least part of the second image, based on the second image being flagged, from the data set;

prepare the modified data set as a series of feature vectors having a length of at least m*n, where m is a number of spectral values included in each of said images in the modified data set, and n is number of image dates which includes said images for the one or more plots in the modified data set, the spectral values including spectral color data from said images and NDVI values;

access phenotypic data for a crop included in the one or more plots, at a time consistent with the images associated with the one or more plots, the phenotypic data including relative maturity;

train a model, based on data including the series of feature vectors for the one or more plots, and the accessed phenotypic data, to identify relative maturity of a target crop in a target plot from images of the target plot, whereby weights of the model are fitted to said data; and

store the trained model in memory.

9 . The system of claim 8 , wherein the computing device is configured, in order to compare the first NDVI value of the first image of the plot at time n to the second NDVI value of the second image of the plot at time n+1, to compare each NDVI value of the first image to corresponding NDVI values of the second image, where the NDVI values correspond based on a corresponding location in the plot; and

wherein the computing device is configured, in order to flag the second image in response to the second NDVI value being greater than the first NDVI value, to flag the second image in response to each NDVI value of the second image being greater than each corresponding NDVI value of the first image.

10 . The system of claim 8 , wherein the model is one of: a random forest algorithm and a boosting algorithm.

11 . The system of claim 8 , wherein the series of feature vectors further includes a NDVI delta, wherein the NDVI delta includes a difference between the NDVI values of consecutive images in the data set.

12 . The system of claim 8 , wherein the computing device is further configured to forecast phenotypic data for a plot based on the trained model.

13 . The system of claim 8 , wherein the computing device is further configured to generate, via the model, phenotypic data for the crop planted in at least one of the one or more plots.

14 . The system of claim 8 , wherein the computing device is further configured to:

calculate a confidence interval for the model based on the generated phenotypic data for the crop planted in the at least one of the one or more plots; and

store the calculated confidence interval in the memory in association with the model.

15 . A non-transitory computer-readable storage medium including executable instructions for processing image data, which when executed by at least one processor, cause the at least one processor to:

access a data set, the data set including images associated with one or more plots, the images having a spatial resolution of better than or equal to one foot by one foot per pixel;

for each of the one or more plots:

compare, for each pixel in the images, a first normalized difference vegetation index (NDVI) value of a first image of the plot at time n from the data set to a second NDVI value of a second image of the plot at time n+1 from the data set;

determine differences of the second NDVI values relative to the first NDVI values;

in response to a first threshold percentage of the differences being positive, flag at least part of the second image; and

modify the data set by removing the at least part of the second image, based on the second image being flagged, from the data set;

prepare the modified data set as a series of feature vectors having a length of at least m*n, where m is a number of spectral values included in each of said images in the modified data set, and n is number of image dates which includes said images for the one or more plots in the modified data set, the spectral values including spectral color data from said images and NDVI values;

access phenotypic data for a crop included in the one or more plots, at a time consistent with the images associated with the one or more plots, the phenotypic data including relative maturity;

train a model, based on data including the series of feature vectors for the one or more plots, and the accessed phenotypic data, to identify relative maturity of a target crop in a target plot from images of the target plot, whereby weights of the model are fitted to said data; and

store the trained model in memory.