Autonomous detection and control of vegetation
A method includes obtaining, by the treatment system configured to implement a machine learning (ML) algorithm, one or more images of a region of an agricultural environment near the treatment system, wherein the one or more images are captured from the region of a real-world where agricultural target objects are expected to be present, determining one or more parameters for use with the ML algorithm, wherein at least one of the one or more parameters is based on one or more ML models related to identification of an agricultural object, determining a real-world target in the one or more images using the ML algorithm, wherein the ML algorithm is at least partly implemented using the one or more processors of the treatment system, and applying a treatment to the target by selectively activating the treatment mechanism based on a result of the determining the target.
1. A method implemented by a treatment system having one or more processor, a storage, and a treatment mechanism, comprising:
obtaining, by the treatment system mountable on an agricultural vehicle and configured to implement a machine learning (ML) algorithm, one or more images of a region of an agricultural environment near the treatment system, wherein the one or more images are captured from the region of a real-world where agricultural target objects are expected to be present;
determining, by the treatment system, one or more parameters for use with the ML algorithm, wherein at least one of the one or more parameters is based on one or more ML models related to identification of an agricultural object;
determining, by the treatment system, a real-world target in the one or more images using the ML algorithm, wherein the ML algorithm is at least partly implemented using the one or more processors of the treatment system; and
applying a treatment to the real-world target by selectively activating the treatment mechanism based on a result of the determining the target;
wherein the determining the target includes:
identifying a subset of pixels of the one or more images for a next processing;
detecting, using the ML algorithm, one or more areas in the subset of pixels including an area that includes the target; and
performing refinement of the one or more areas.
2. The method of claim 1 , wherein the identifying is performed using a first computer vision technique and the refinement is performed using a second computer vision technique.
3. The method of claim 2 , wherein the refinement comprises identifying a foreground object and a background in the one or more areas.
4. The method of claim 3 , wherein the identifying is performed by segmenting according to colors or by segmenting according to regions or by segmenting according to detected edges.
5. The method of claim 2 , wherein the refinement comprises identifying superpixels in the one or more areas.
6. The method of claim 1 , wherein the activating the treatment mechanism includes emitting a fluid projectile towards the target.
7. The method of claim 1 , wherein the activating the treatment mechanism includes orienting the treatment mechanism towards the target and emitting a beam of light towards the target.
8. A method implemented by a treatment system having one or more processor, a storage, and a treatment mechanism, comprising:
obtaining, by the treatment system mountable on an agricultural vehicle and configured to implement a machine learning (ML) algorithm, one or more images of a region of an agricultural environment near the treatment system, wherein the one or more images are captured from the region of a real-world where agricultural target objects are expected to be present;
determining, by the treatment system, one or more parameters for use with the ML algorithm, wherein at least one of the one or more parameters is based on one or more ML models related to identification of an agricultural object;
determining, by the treatment system, a real-world target in the one or more images using the ML algorithm, wherein the ML algorithm is at least partly implemented using the one or more processors of the treatment system; and
applying a treatment to the real-world target by selectively activating the treatment mechanism based on a result of the determining the target;
wherein the selectively activating comprises activating the treatment mechanism in response to detecting the target in multiple consecutive frames of the one or more images.
9. The method of claim 8 wherein the ML algorithm is implemented using a convolutional neural network (CNN).
10. A treatment system comprising one or more processors, a storage, and a treatment mechanism, wherein the one or more processors are configured to:
obtaining one or more images of a region of an agricultural environment near the treatment system, wherein the one or more images are captured from the region of a real-world where agricultural target objects are expected to be present;
determine one or more parameters for use with the ML algorithm, wherein at least one of the one or more parameters is based on one or more ML models related to identification of an agricultural object;
determine a real-world target in the one or more images using the ML algorithm, wherein the ML algorithm is at least partly implemented using the one or more processors of the treatment system; and
apply a treatment to the target by selectively activating the treatment mechanism based on a result of the determining the target;
wherein the real-world target is determined by:
identifying a subset of pixels of the one or more images for a next processing;
detecting, using the ML algorithm, one or more areas in the subset of pixels including an area that includes the target; and
performing refinement of the one or more areas.
11. The treatment system of claim 10 , wherein the identifying is performed using a first computer vision technique and the refinement is performed using a second computer vision technique.
12. The treatment system of claim 11 , wherein the refinement comprises identifying a foreground object and a background in the one or more areas.
13. The treatment system of claim 12 , wherein the identifying is performed by segmenting according to colors or by segmenting according to regions or by segmenting according to detected edges.
14. The treatment system of claim 11 , wherein the refinement comprises identifying superpixels in the one or more areas.
15. The treatment system of claim 10 , wherein the activating the treatment mechanism includes emitting a fluid projectile towards the target.
16. The treatment system of claim 10 , wherein the selectively activating the treatment mechanism includes orienting the treatment mechanism towards the target and emitting a beam of light towards the target.
17. The treatment system of claim 10 , wherein the selectively activating comprises activating the treatment mechanism in response to detecting the target in multiple consecutive frames of the one or more images.
18. The treatment system of claim 10 , wherein the ML algorithm is implemented using a convolutional neural network (CNN).
19. A non-transitory computer-readable storage medium having code stored thereon, the code, upon execution, causing one or more processors of a treatment system mountable on an agricultural vehicle to implement a method, comprising:
obtaining, by the treatment system, one or more images of a region of an agricultural environment near the treatment system, wherein the one or more images are captured from the region of a real-world where agricultural target objects are expected to be present;
determining, by the treatment system, one or more parameters for use with a machine learning (ML) algorithm, wherein at least one of the one or more parameters is based on one or more ML models related to identification of an agricultural object;
determining, by the treatment system, a real-world target in the one or more images using the ML algorithm, wherein the ML algorithm is at least partly implemented using the one or more processors of the treatment system; and
applying a treatment to the target by selectively activating the treatment mechanism based on a result of the determining the target;
wherein the selectively activating comprises activating the treatment mechanism in response to detecting the target in multiple consecutive frames of the one or more images.
20. The non-transitory computer-readable storage medium of claim 19 , wherein the determining the target includes:
identifying a subset of pixels of the one or more images for a next processing;
detecting, using the ML algorithm, one or more areas in the subset of pixels including an area that includes the target; and
performing refinement of the one or more areas.
21. The non-transitory computer-readable storage medium of claim 20 , wherein the identifying is performed using a first computer vision technique and the refinement is performed using a second computer vision technique.
22. The non-transitory computer-readable storage medium of claim 21 , wherein the refinement comprises identifying a foreground object and a background in the one or more areas.
23. The non-transitory computer-readable storage medium of claim 22 , wherein the identifying is performed by segmenting according to colors or by segmenting according to regions or by segmenting according to detected edges.
24. The non-transitory computer-readable storage medium of claim 21 , wherein the refinement comprises identifying superpixels in the one or more areas.
25. The non-transitory computer-readable storage medium of claim 19 , wherein the activating the treatment mechanism includes emitting a fluid projectile towards the target.
26. The non-transitory computer-readable storage medium of claim 19 , wherein the activating the treatment mechanism includes orienting the treatment mechanism towards the target and emitting a beam of light towards the target.
27. The non-transitory computer-readable storage medium of claim 19 , wherein the ML algorithm is implemented using a convolutional neural network (CNN).