Modeling and decision support for horticulture
Implementations include providing a baseline multi-dimensional model of a cultivar, determining an encoding based on the baseline multi-dimensional model, and a target multi-dimensional model, the encoding defining a string of symbols, and being based on an alphabet and a set of rules, providing an expected multi-dimensional model based on the encoding, and a modified set of rules, the modified set of rules being based on the set of rules, the expected multi-dimensional model representing the cultivar after a period of time, selecting a set of actions by determining multiple predicted multi-dimensional models for each set of actions in a plurality of sets of actions, and, for each predicted multi-dimensional model, providing a predicted yield that can be used to determine impact with respect an expected yield, the set of actions being selected based on a respective impact, and providing the set of actions as output for executing on the cultivar.
1. A computer-implemented method executed by one or more processors, the method comprising:
providing a baseline multi-dimensional data model of a cultivar, the cultivar growing over time;
determining an encoding based on the baseline multi-dimensional data model, and a target multi-dimensional model, the encoding defining a string of symbols, and being based on an alphabet and a set of rules;
providing an expected multi-dimensional model based on the encoding, and a modified set of rules, the modified set of rules being based on the set of rules, the expected multi-dimensional model representing the cultivar after a period of time;
selecting a set of actions by determining multiple predicted multi-dimensional models for each set of actions in a plurality of sets of actions, and, for each predicted multi-dimensional model, providing a predicted yield that can be used to determine impact with respect an expected yield, the set of actions being selected based on a respective impact; and
providing the set of actions as output for executing one or more actions of the set of actions on the cultivar.
2. The method of claim 1 , further comprising receiving image data representative of stereoscopic images of the cultivar, wherein the baseline multi-dimensional data model is generated at least partially from the stereoscopic images.
3. The method of claim 1 , wherein the alphabet and set of rules are provided as part of an L-system for providing strings to represent the cultivar.
4. The method of claim 1 , wherein the modified set of rules are provided by adjusting a probability of at least one symbol of at least one rule based on the context data.
5. The method of claim 1 , wherein the target multi-dimensional model is selected from a plurality of generated target multi-dimensional models based on respective similarity scores, each similarity score representing a degree of similarity between the baseline multi-dimensional model, and a respective target multi-dimensional model.
6. The method of claim 5 , wherein the plurality of target multi-dimensional models are generated based on one of a recurrent neural network (RNN) approach, and an estimation of distribution algorithm (EDA) approach.
7. The method of claim 1 , wherein the set of actions is output by displaying the cultivar on a display, and providing one or more treatments, each treatment representing an action in the set of actions.
8. The method of claim 1 , wherein at least one action in the set of actions comprises pruning a feature of the cultivar.
9. The method of claim 8 , wherein the feature comprises one or more of a leaf and a branch.
10. The method of claim 1 , wherein each rule in the set of rules provides two or more expansion symbols for replacing symbols in previous iterations of the string of symbols.
11. The method of claim 1 , further comprising receiving environmental sensor data, wherein the expected multi-dimensional model is provided based on the received environmental sensor data.
12. The method of claim 1 , wherein each predicted multi-dimensional model is determined based on a set of actions applied to the expected multi-dimensional model.
13. A non-transitory computer-readable storage medium coupled to one or more processors and having instructions stored thereon which, when executed by the one or more processors, cause the one or more processors to perform operations comprising:
providing a baseline multi-dimensional data model of a cultivar, the cultivar growing over time;
determining an encoding based on the baseline multi-dimensional data model, and a target multi-dimensional model, the encoding defining a string of symbols, and being based on an alphabet and a set of rules;
providing an expected multi-dimensional model based on the encoding, and a modified set of rules, the modified set of rules being based on the set of rules, the expected multi-dimensional model representing the cultivar after a period of time;
selecting a set of actions by determining multiple predicted multi-dimensional models for each set of actions in a plurality of sets of actions, and, for each predicted multi-dimensional model, providing a predicted yield that can be used to determine impact with respect an expected yield, the set of actions being selected based on a respective impact; and
providing the set of actions as output for executing one or more actions of the set of actions on the cultivar.
14. The computer-readable storage medium of claim 13 , wherein operations further comprise receiving image data representative of stereoscopic images of the cultivar, wherein the baseline multi-dimensional data model is generated at least partially from the stereoscopic images.
15. The computer-readable storage medium of claim 13 , wherein the alphabet and set of rules are provided as part of an L-system for providing strings to represent the cultivar.
16. The computer-readable storage medium of claim 13 , wherein the modified set of rules are provided by adjusting a probability of at least one symbol of at least one rule based on the context data.
17. The computer-readable storage medium of claim 13 , wherein the target multi-dimensional model is selected from a plurality of generated target multi-dimensional models based on respective similarity scores, each similarity score representing a degree of similarity between the baseline multi-dimensional model, and a respective target multi-dimensional model.
18. The computer-readable storage medium of claim 17 , wherein the plurality of target multi-dimensional models are generated based on one of a recurrent neural network (RNN) approach, and an estimation of distribution algorithm (EDA) approach.
19. The computer-readable storage medium of claim 13 , wherein the set of actions is output by displaying the cultivar on a display, and providing one or more treatments, each treatment representing an action in the set of actions.
20. The computer-readable storage medium of claim 13 , wherein at least one action in the set of actions comprises pruning a feature of the cultivar.
21. The computer-readable storage medium of claim 20 , wherein the feature comprises one or more of a leaf and a branch.
22. The computer-readable storage medium of claim 13 , wherein each rule in the set of rules provides two or more expansion symbols for replacing symbols in previous iterations of the string of symbols.
23. The computer-readable storage medium of claim 13 , wherein operations further comprise receiving environmental sensor data, wherein the expected multi-dimensional model is provided based on the received environmental sensor data.
24. The computer-readable storage medium of claim 13 , wherein each predicted multi-dimensional model is determined based on a set of actions applied to the expected multi-dimensional model.
25. A system, comprising:
one or more processors; and
a computer-readable storage device coupled to the one or more processors and having instructions stored thereon which, when executed by the one or more processors, cause the one or more processors to perform operations comprising:
providing a baseline multi-dimensional data model of a cultivar, the cultivar growing over time;
determining an encoding based on the baseline multi-dimensional data model, and a target multi-dimensional model, the encoding defining a string of symbols, and being based on an alphabet and a set of rules;
providing an expected multi-dimensional model based on the encoding, and a modified set of rules, the modified set of rules being based on the set of rules, the expected multi-dimensional model representing the cultivar after a period of time;
selecting a set of actions by determining multiple predicted multi-dimensional models for each set of actions in a plurality of sets of actions, and, for each predicted multi-dimensional model, providing a predicted yield that can be used to determine impact with respect an expected yield, the set of actions being selected based on a respective impact; and
providing the set of actions as output for executing one or more actions of the set of actions on the cultivar.
26. The system of claim 25 , wherein operations further comprise receiving image data representative of stereoscopic images of the cultivar, wherein the baseline multi-dimensional data model is generated at least partially from the stereoscopic images.
27. The system of claim 25 , wherein the alphabet and set of rules are provided as part of an L-system for providing strings to represent the cultivar.
28. The system of claim 25 , wherein the modified set of rules are provided by adjusting a probability of at least one symbol of at least one rule based on the context data.
29. The system of claim 25 , wherein the target multi-dimensional model is selected from a plurality of generated target multi-dimensional models based on respective similarity scores, each similarity score representing a degree of similarity between the baseline multi-dimensional model, and a respective target multi-dimensional model.
30. The system medium of claim 29 , wherein the plurality of target multi-dimensional models are generated based on one of a recurrent neural network (RNN) approach, and an estimation of distribution algorithm (EDA) approach.
31. The system of claim 25 , wherein the set of actions is output by displaying the cultivar on a display, and providing one or more treatments, each treatment representing an action in the set of actions.
32. The system of claim 25 , wherein at least one action in the set of actions comprises pruning a feature of the cultivar.
33. The system of claim 32 , wherein the feature comprises one or more of a leaf and a branch.
34. The system of claim 25 , wherein each rule in the set of rules provides two or more expansion symbols for replacing symbols in previous iterations of the string of symbols.
35. The system of claim 25 , wherein operations further comprise receiving environmental sensor data, wherein the expected multi-dimensional model is provided based on the received environmental sensor data.
36. The system of claim 25 , wherein each predicted multi-dimensional model is determined based on a set of actions applied to the expected multi-dimensional model.