Mapping Field Anomalies Using Digital Images and Machine Learning Models
A computer-implemented method for generating an improved map of field anomalies using digital images and machine learning models is disclosed. In an embodiment, a method comprises: obtaining a shapefile that defines boundaries of an agricultural plot and boundaries of the field containing the plot; obtaining a plurality of plot images within the field from one or more image capturing devices that are located within the boundaries of the field; calibrating and pre-processing the plurality of plot images to create a plot map of the agricultural plot at a plot level; based on the plot map of the agricultural plot, generating a plot grid; based on the plot grid and the plot map, generating a plurality of plot tiles; based on the plurality of plot tiles, generating, using a first machine learning model and a plurality of first image classifiers corresponding to one or more first anomalies, a set of classified plot images that depicts at least one anomaly; based on the set of classified plot images, generating a plot anomaly map for the agricultural plot; transmitting the plot anomaly map to one or more controllers that control one or more agricultural machines or database systems to perform agricultural functions on the agricultural plot.
1 . (canceled)
2 . A system for mapping field anomalies in an agricultural field, the system comprising:
one or more cameras configured to collect one or more above ground images of the agricultural field;
a calibration unit configured to receive and calibrate the one or more images;
a stitcher configured to receive the calibrated images and perform a stitching process on the calibrated images to generate a field level image;
a grid map generator configured to receive the field level image and divide the field level image into a field grid;
a tile generator configured to receive the grid map and generate a plurality of small tiles;
a classification unit comprising a machine learning model and configured to receive the plurality of small tiles and classify each of the plurality of small tiles, and further configured to generate at least one anomaly map including at least a portion of the classified tiles; and
a shapefile generator configured to receive the at least one anomaly map and generate a shapefile comprising geographical coordinates, wherein the geographical coordinates reference the classified tiles.
3 . The system of claim 2 , wherein at least one of the one or more cameras are attached to an aerial vehicle.
4 . The system of claim 2 , wherein at least one of the one or more cameras are mounted on a ground based apparatus.
5 . The system of claim 2 , wherein the machine learning machine learning model is based on at least one of a fusion of classifier output and vegetative index data.
6 . The system of claim 2 , wherein the calibration unit is configured to calibrate the one or more images based on at least one of a color correction, a hue correction, a resolution correction, and a gamma color correction.
7 . The system of claim 2 , wherein the field level image covers a field having an area of between 40 and 100 acres.
8 . The system of claim 2 , wherein the field level image is an orthomosaic image
9 . The system of claim 2 , wherein the anomaly map further includes a legend describing different colors assigned to classified regions.
10 . A method for mapping field anomalies in an agricultural field, the method comprising:
receiving one or more above ground images of the agricultural field from one or more cameras;
calibrating the one or more images;
stitching the calibrated images to generate a field level image;
generating a grid map based on the field level image;
generating a plurality of small tiles based on the grid map;
classifying, using a machine learning model, each of the plurality of small tiles of the grid map,
generating at least one anomaly map including at least a portion of the classified tiles; and
generating a shapefile including the at least one anomaly map and further comprising geographical coordinates, wherein the geographical coordinates reference the classified tiles.
11 . The method of claim 10 , wherein at least one of the one or more cameras are attached to an aerial vehicle.
12 . The method of claim 10 , wherein at least one of the one or more cameras are mounted on a ground based apparatus.
13 . The method of claim 10 , wherein the machine learning machine learning model is based on at least one of a fusion of classifier output and vegetative index data.
14 . The method of claim 10 , wherein the one or more images are calibrated based on at least one of a color correction, a hue correction, a resolution correction, and a gamma color correction.
15 . The method of claim 10 , wherein the field level image covers a field having an area of between 40 and 100 acres.
16 . The method of claim 10 , wherein the field level image is an orthomosaic image
17 . The method of claim 10 , wherein the anomaly map further includes a legend describing different colors assigned to classified regions.
18 . A non-transitory computer readable storage medium comprising:
at least one memory section configured to store operational instructions that, when executed by one or more processing modules of one or more computing devices affiliated with agricultural equipment, cause the one or more computing devices to:
receiving one or more above ground images of the agricultural field from one or more cameras;
calibrate the one or more images;
stitch the calibrated images to generate a field level image;
generate a grid map based on the field level image;
generate a plurality of small tiles based on the grid map;
classify, using a machine learning model, each of the plurality of small tiles of the grid map,
generate at least one anomaly map including at least a portion of the classified tiles; and
generate a shapefile including the at least one anomaly map and further comprising geographical coordinates, wherein the geographical coordinates reference the classified tiles.
19 . The non-transitory computer readable storage medium of claim 18 , wherein at least one of the one or more cameras are attached to an aerial vehicle.
20 . The non-transitory computer readable storage medium of claim 18 , wherein the machine learning machine learning model is based on at least one of a fusion of classifier output and vegetative index data.
21 . The non-transitory computer readable storage medium of claim 18 , wherein the one or more images are calibrated based on at least one of a color correction, a hue correction, a resolution correction, and a gamma color correction.