Machine learning method and system for predicting key agricultural field management practices

A non-transitory computer-readable storage medium storing instructions that, when executed by a computer, cause the computer to perform a method for predicting agricultural management practices, the method comprising:

• generating a training dataset that comprises a plurality of years of known management practices associated with a plurality of fields dispersed within geographic region along with a corresponding plurality of years of first remote sense images;
• training a Bayesian crop model to predict the plurality of years of known management practices associated with the plurality of fields using the corresponding plurality of years of first remote sense images as inputs, the training comprising
• cleansing the corresponding plurality of years of remote sense images by:
• removing images with first missing data below a prescribed threshold;
• retaining images with second missing data above the prescribed threshold;
• time-processing third data from other time-adjacent images that corresponds to the second missing data;
• and
• using the time-processed third data to replace the second missing data;
• combining spectral band images for each observation date to generate corresponding first composite images
• combining first composite images for each observation date to create composite estimated growth curves for each of the plurality of fields;
• wherein the time-processing of the third data comprises generating reconstructed spectral values for missing portions of the remote sense images using temporally adjacent remote sensing images, and wherein combining the spectral band images comprises generating composite multi-spectral images based on pixel-level spectral intensity values;
• and
• employing generative Bayesian inference techniques that match the composite estimated growth curves to corresponding known crop vegetation index double-sigmoid functions to estimate first associated posterior distribution parameters;
• providing a time series of second remote sense images associated with a corresponding field having unknown management practices as exclusive inputs to the Bayesian crop model;
• and
• executing the Bayesian crop model to predict a crop type, a planting date, an emergence date, a senescence start date, and a harvest date for the corresponding field, the executing comprising:
• treating the estimated first associated posterior distribution parameters as known quantities;
• and
• using the known quantities to generate the crop type, the planting date, the emergence date, the senescence start date, and the harvest date from estimated second associated posterior distribution parameters derived from the time series of second remote sense images.

A system for predicting agricultural management practices, the system comprising:

• an incentive program server, the server comprising:
• a crop model processor, comprising a Bayesian crop model;
• and
• a training processor, configured to:
• generate a training dataset that comprises a plurality of years of known management practices associated with a plurality of fields dispersed within geographic region along with a corresponding plurality of years of first remote sense images;
• and
• train the Bayesian crop model to predict the plurality of years of known management practices associated with the plurality of fields using the corresponding plurality of years of first remote sense images as inputs, wherein training the Bayesian crop model comprises:
• cleansing the corresponding plurality of years of remote sense images by:
• removing images with first missing data below a prescribed threshold;
• retaining images with second missing data above the prescribed threshold;
• time-processing third data from other time-adjacent images that corresponds to the second missing data;
• and
• using the time-processed third data to replace the second missing data;
• combining spectral band images for each observation date to generate corresponding first composite images;
• combining first composite images for each observation date to create composite estimated growth curves for each of the plurality of fields,
• wherein the time-processing of the third data comprises generating reconstructed spectral values for missing portions of the remote sense images using temporally adjacent remote sensing images, and wherein combining the spectral band images comprises generating composite multi-spectral images based on pixel-level spectral intensity values, and
• employing generative Bayesian inference techniques that match the composite estimated growth curves to corresponding known crop vegetation index double-sigmoid functions to estimate first associated posterior distribution parameters;
• a remote sense processor, configured to provide a time series of second remote sense images associated with a corresponding field having unknown management practices as exclusive inputs to the Bayesian crop model within the crop model processor;
• the crop model processor, configured to execute the Bayesian crop model to predict a crop type, a planting date, an emergence date, a senescence start date, and a harvest date for the corresponding field, wherein executing the Bayesian crop model comprises:
• treating the estimated first associated posterior distribution parameters as known quantities;
• and
• using the known quantities to generate the crop type, the planting date, the emergence date, the senescence start date, and the harvest date from estimated second associated posterior distribution parameters derived from the time series of second remote sense images.