Detecting and tracking features of plants for treatment in an agricultural environment
A method and system having instructions to perform actions include obtaining images of an agricultural environment including a first image comprising at least one background portion and one or more regions of interest, implementing a machine learning (ML) algorithm on a portion of the first image including a portion of the background portion and the one or more regions of interest, detecting a plurality of objects associated with a plurality of real-world objects in the agricultural environment in at least one region of interest in the one or more regions of interest of the first image including detecting a first object and detecting a second object, implementing a second algorithm on the portion of the first image comprising the first object to detect one or more divided features of the first object, tracking a feature of the one or more divided features across subsequent images as the moving platform traverses the agricultural environment, tracking the second object, and selecting a target action configured to target the second object and applying the target action via the target mechanism to the second object.
1. A method implemented by a treatment system couple to a moving platform, the treatment system having one or more processors, a storage, and a target mechanism, comprising:
obtaining images of an agricultural environment including a first image comprising at least one background portion and one or more regions of interest;
implementing a machine learning (ML) algorithm on a portion of the first image including a portion of the background portion and the one or more regions of interest;
detecting a plurality of objects associated with a plurality of real-world objects in the agricultural environment in at least one region of interest in the one or more regions of interest of the first image including detecting a first object and detecting a second object;
implementing a second algorithm on the portion of the first image comprising the first object to detect one or more divided features of the first object;
tracking a feature of the one or more divided features across subsequent images as the moving platform traverses the agricultural environment;
tracking the second object; and
selecting a target action configured to target the second object and applying the target action via the target mechanism to the second object.
2. The method of claim 1 , further comprising generating patch of the first image containing the first object, second object, or both.
3. The method of claim 2 , further comprising tracking the patch across subsequent images as the moving platform traverses the agricultural environment.
4. The method of claim 1 , further comprising identifying the background portion as a brown background, green background, the background portion comprising a plurality of landmarks, or a combination thereof.
5. The method of claim 4 , further comprising selecting the ML algorithm from a plurality of ML algorithms, each ML algorithm from the plurality of ML algorithms configured to at least detect and classify plant objects or features of plant objects.
6. The method of claim 5 , wherein selecting the ML algorithm from the plurality of ML algorithms is based on the identified background portion.
7. The method of claim 1 , wherein the second algorithm is a second ML algorithm configured to detect and classify plant object, portions of plant object, features of plant objects, or a combination thereof.
8. The method of claim 7 , wherein the features of plant objects comprise a root, a stem or a portion of a stem, a bud, a portion of a root, a portion of a plant surface, a landmark, a phenological attribute, or a combination thereof.
9. The method of claim 7 , wherein the one or more divided features of the first object comprises one or more pixel or cluster of pixels of the first images associated with the one or more divided features.
10. The method of claim 9 , wherein tracking the feature of the one or more divided features across subsequently obtained images comprises tracking the one or more pixels or cluster of pixels associated with the feature.
11. The method of claim 1 , further comprising refraining from or cancelling applying the target action to the second object upon determining that a portion of the first object, the portion associated with the detected feature of one or more divided features of the first object, is within a threshold proximity to the second object.
12. A non-transitory computer-readable medium comprising instructions that, when executed by one or more processors of an agricultural treatment system, cause the agricultural treatment system and a treatment mechanism coupled to the agricultural treatment system to perform operations comprising:
obtain images of an agricultural environment including a first image comprising at least one background portion and one or more regions of interest;
implement a machine learning (ML) algorithm on a portion of the first image including a portion of the background portion and the one or more regions of interest;
detect a plurality of objects associated with a plurality of real-world objects in the agricultural environment in at least one region of interest in the one or more regions of interest of the first image including detecting a first object and detecting a second object;
implement a second algorithm on the portion of the first image comprising the first object to detect one or more divided features of the first object;
track a feature of the one or more divided features across subsequent images as the agricultural treatment system traverses the agricultural environment;
track the second object; and
select a target action configured to target the second object and applying the target action via the target mechanism to the second object.
13. The non-transitory computer-readable medium of claim 12 , further comprising generate patch of the first image containing the first object, second object, or both.
14. The non-transitory computer-readable medium of claim 13 , further comprising track the patch across subsequent images as the agricultural treatment system traverses the agricultural environment.
15. The non-transitory computer-readable medium of claim 12 , further comprising identify the background portion as a brown background, green background, the background portion comprising a plurality of landmarks, or a combination thereof.
16. The non-transitory computer-readable medium of claim 15 , further comprising select the ML algorithm from a plurality of ML algorithms, each ML algorithm from the plurality of ML algorithms configured to at least detect and classify plant objects or features of plant objects.
17. The non-transitory computer-readable medium of claim 16 , wherein selecting the ML algorithm from the plurality of ML algorithms is based on identified background portion.
18. The non-transitory computer-readable medium of claim 12 , wherein the second algorithm is a second ML algorithm configured to detect and classify plant object, portions of plant object, features of plant objects, or a combination thereof.
19. The non-transitory computer-readable medium of claim 18 , wherein the features of plant objects comprise a root, a portion of a root, a portion of a surface, a stem or a portion of a stem, a bud, a landmark, a phenological attribute, or a combination thereof.
20. The non-transitory computer-readable medium of claim 18 , wherein the one or more divided features of the first object comprises one or more pixel or cluster of pixels of the first images associated with the one or more divided features.
21. The non-transitory computer-readable medium of claim 20 , wherein tracking the feature of the one or more divided features across subsequently obtained images comprises tracking the one or more pixels or cluster of pixels associated with the feature.
22. The non-transitory computer-readable medium of claim 12 , further comprising refrain from or cancel applying the target action to the second object upon determining that a portion of the first object, the portion associated with the detected feature of one or more divided features of the first object, is within a threshold proximity to the second object.