Non-transitory computer-readable media and devices for blade wear monitoring
Provided is a non-transitory computer readable medium storing instructions that, when executed by at least one processor, cause the at least one processor to obtain a first signal based on an input image using a trained machine learning model, the input image being an image of a plant cut by a blade, and the first signal indicating a wear level of the blade, determine whether a level of the first signal is greater than or equal to a threshold, generate a second signal in response to determining the level of the first signal is greater than or equal to the threshold, and output the second signal.
1. A non-transitory computer-readable medium storing instructions that, when executed by at least one processor, cause the at least one processor to:
obtain a first signal using a trained machine learning model, the first signal being based on an input image, the input image being an image of a plant cut by a blade, the input image being captured during a harvesting operation in which the plant is cut by the blade, and the first signal indicating a wear level of the blade;
determine whether a level of the first signal is greater than or equal to a threshold without stopping the harvesting operation; and
control a rotational speed of the blade in response to determining that the level of the first signal is greater than or equal to the threshold, the control of the rotational speed of the blade being performed without stopping the harvesting operation.
2. The non-transitory computer-readable medium of claim 1 , wherein the trained machine learning model was trained using a plurality of reference images, each of the plurality of reference images being an image of a respective plant cut by a respective blade, and each of the plurality of reference images being associated with a respective blade wear level.
3. The non-transitory computer-readable medium of claim 2 , wherein the respective blade wear level is based on one of a billet length of the respective plant, a cleanness of the cut of the respective plant or a measure of dullness of the respective blade.
4. The non-transitory computer-readable medium of claim 1 , wherein the instructions, when executed by the at least one processor, cause the at least one processor to:
obtain an ordered set of images captured of plants cut by a first blade, a first image among the set of images corresponding to a lowest blade wear level of the first blade, and a last image among the set of images corresponding to a blade wear level of the first blade exceeding the threshold; and
train a machine learning model based on the set of images to obtain the trained machine learning model.
5. The non-transitory computer-readable medium of claim 1 , wherein the plant is a billet of sugar cane.
6. The non-transitory computer-readable medium of claim 1 , wherein
the threshold is one of a plurality of thresholds, and each respective threshold among the plurality of thresholds corresponds to a respective alert signal among a plurality of alert signals; and
the instructions, when executed by the at least one processor, cause the at least one processor to,
generate the respective alert signal in response to determining the first signal is greater than or equal to the respective threshold corresponding to the respective alert signal, and
output the respective alert signal.
7. The non-transitory computer-readable medium of claim 6 , wherein the instructions, when executed by the at least one processor, cause the at least one processor to:
output the respective alert signal to a server; and
receive the plurality of thresholds from the server.
8. The non-transitory computer-readable medium of claim 1 , wherein the instructions, when executed by the at least one processor, cause the at least one processor to receive the trained machine learning model from an external source.
9. The non-transitory computer-readable medium of claim 1 , wherein the instructions, when executed by the at least one processor, cause the at least one processor to:
output the input image to a server;
receive a re-trained machine learning model from the server; and
obtain a third signal using the re-trained machine learning model.
10. The non-transitory computer-readable medium of claim 1 , wherein the instructions, when executed by the at least one processor, cause the at least one processor to:
re-train the trained machine learning model based on the input image to obtain a re-trained machine learning model; and
obtain a third signal using the re-trained machine learning model.
11. The non-transitory computer-readable medium of claim 1 , wherein the instructions, when executed by the at least one processor, cause the at least one processor to reduce a ground speed of a vehicle in response to determining the level of the first signal is greater than or equal to the threshold.
12. The non-transitory computer-readable medium of claim 1 , wherein the instructions, when executed by the at least one processor, cause the at least one processor to increase the rotational speed of the blade in response to determining the level of the first signal is greater than or equal to the threshold.
13. The non-transitory computer-readable medium of claim 1 , wherein the instructions, when executed by the at least one processor, cause the at least one processor to:
train a machine learning model using a plurality of reference images to obtain the trained machine learning model, each of the plurality of reference images being an image of a respective plant cut by a blade;
re-train the trained machine learning model based on the input image to obtain a re-trained machine learning model; and
obtain a third signal using the re-trained machine learning model.
14. A device for blade wear monitoring, the device comprising:
processing circuitry configured to,
obtain a first signal based on an input image using a trained machine learning model, the input image being an image of a plant cut by a blade, the input image being captured during a harvesting operation in which the plant is cut by the blade, and the first signal indicating a wear level of the blade,
determine whether a level of the first signal is greater than or equal to a threshold without stopping the harvesting operation, and
control a rotational speed of the blade in response to determining that the level of the first signal is greater than or equal to the threshold, the control of the rotational speed of the blade being performed without stopping the harvesting operation.
15. The device of claim 14 , wherein the trained machine learning model was trained using a plurality of reference images, each of the plurality of reference images being an image of a respective plant cut by a respective blade, and each of the plurality of reference images being associated with a respective blade wear level.
16. The device of claim 15 , wherein the respective blade wear level is based on one of a billet length of the respective plant, a cleanness of the cut of the respective plant or a measure of dullness of the respective blade.
17. The device of claim 14 , wherein the processing circuitry is configured to:
obtain an ordered set of images captured of plants cut by a first blade, a first image among the set of images corresponding to a lowest blade wear level of the first blade, and a last image among the set of images corresponding to a blade wear level of the first blade exceeding the threshold; and
train a machine learning model based on the set of images to obtain the trained machine learning model.
18. The device of claim 14 , wherein the plant is a billet of sugar cane.
19. The device of claim 14 , wherein
the threshold is one of a plurality of thresholds, and each respective threshold among the plurality of thresholds corresponds to a respective alert signal among a plurality of alert signals; and
the processing circuitry is configured to,
generate the respective alert signal in response to determining the first signal is greater than or equal to the respective threshold corresponding to the respective alert signal,
output the respective alert signal to a server, and
receive the plurality of thresholds from the server.
20. The device of claim 14 , wherein the processing circuitry is configured to:
output the input image to a server;
receive a re-trained machine learning model from the server; and
obtain a third signal using the re-trained machine learning model.
21. The device of claim 14 , wherein the processing circuitry is configured to:
re-train the trained machine learning model based on the input image to obtain a re-trained machine learning model; and
obtain a third signal using the re-trained machine learning model.
22. The device of claim 14 , wherein the processing circuitry is configured to reduce a ground speed of a vehicle in response to determining the level of the first signal is greater than or equal to the threshold.
23. The device of claim 14 , wherein the processing circuitry is configured to increase the rotational speed of the blade in response to determining the level of the first signal is greater than or equal to the threshold.