Imaging system with unsupervised learning
An imaging system and method uses grouping and elimination to label images of unknown items. The items may be stacked together with known or unknown items. The items may be packages, such as packages of beverage containers. A machine learning model may be used to infer skus of the packages. The machine learning model is trained on known skus but is not trained on unknown skus. Multiple images of the same unknown sku are grouped using the machine learning model. Elimination based upon lists of expected skus is used to label each group of unknown skus.
1. A method for identifying a plurality of items in a plurality of stacks using a computing system including at least one machine learning model, the method including:
a) receiving at least one image of the plurality of items in one of the plurality of stacks;
b) analyzing the at least one image using the at least one machine learning model;
c) comparing the plurality of items to a list of expected items expected to be in the one of the plurality of stacks;
d) repeating steps a) to c) for each of the plurality of stacks of the plurality of items;
e) using the at least one machine learning model, forming a plurality of groups of images of the plurality of items in the plurality of stacks; and
f) based upon the lists of expected items, assigning one of the expected items to each of the plurality of groups of images.
2. The method of claim 1 further including:
g) using the plurality of groups of images and the assigned expected items to train the at least one machine learning model.
3. The method of claim 1 further including a step of imaging a plurality of items in the one of the plurality of stacks prior to step a).
4. The method of claim 1 wherein step f) includes using elimination based upon the images of the plurality of items in the stacks.
5. The method of claim 1 wherein the plurality of items in the plurality of stacks are stacked on a plurality of pallets.
6. The method of claim 5 wherein the plurality of items are a plurality of packages.
7. The method of claim 6 wherein the plurality of packages contain beverage containers and wherein the at least one machine learning model is trained on images of a plurality of known packages containing beverage containers.
8. A computing system for evaluating a plurality of items in a plurality of stacks comprising:
at least one processor; and
at least one non-transitory computer-readable medium storing:
at least one machine learning model; and
instructions that, when executed by the at least one processor, cause the computing system to perform operations comprising:
a) receiving at least one image of one of the plurality of stacks of the plurality of items;
b) analyzing the at least one image using the at least one machine learning model;
c) comparing the plurality of items to a list of expected items expected to be in the one of the plurality of stacks;
d) repeating steps a) to c) for each of the plurality of stacks of the plurality of items;
e) using the at least one machine learning model, forming a plurality of groups of images of the plurality of items in of the stacks; and
f) based upon the lists of expected items, assigning one of the expected items to each of the plurality of groups of images.
9. The system of claim 8 wherein the operations further include:
g) using the plurality of groups of images and the assigned expected items to train the at least one machine learning model.
10. The system of claim 8 wherein operation f) includes using elimination based upon the images of the plurality of items in the plurality of stacks.
11. The system of claim 8 wherein the plurality of items in the plurality of stacks are stacked on a plurality of pallets.
12. The system of claim 11 wherein the plurality of items are a plurality of packages.
13. The system of claim 12 wherein the plurality of packages contain beverage containers and wherein the at least one machine learning model is trained on images of a plurality of known packages containing beverage containers.
14. The system of claim 13 further including at least one camera for taking the at least one image of the stack of the plurality of items.
15. The method of claim 1 wherein the at least one image includes an image of each of four sides of the plurality of items in the stack.
16. The system of claim 8 wherein the at least one image includes an image of each of four sides of the plurality of items in the stack.
17. The method of claim 1 wherein the list of expected items in step c) is an order for the items expected to be in the one of the plurality of stacks.
18. The method of claim 1 wherein the list of expected items in step c) is a list of items to be shipped to a store.
19. The method of claim 1 further including the step of: prior to step a) presenting the list of expected items to instruct a user to place the plurality of items in the one of the plurality of stacks.
20. The system of claim 8 wherein the list of expected items in operation c) is an order for the items expected to be in the one of the plurality of stacks.
21. The system of claim 8 wherein the list of expected items in operation c) is a list of items to be shipped to a store.
22. The system of claim 8 further including the operation of: prior to step a) presenting the list of expected items to instruct a user to place the plurality of items in the one of the plurality of stacks.
23. A computing system for evaluating a plurality of packages in a plurality of stacks comprising:
at least one processor; and
at least one non-transitory computer-readable medium storing:
at least one machine learning model trained on images of a plurality of known packages; and
instructions that, when executed by the at least one processor, cause the computing system to perform operations comprising:
a) receiving at least one image of each of a subset of the plurality of packages in one of the plurality of stacks;
b) analyzing the at least one image using the at least one machine learning model to generate an output for each of the subset of the plurality of packages in the one of the plurality of stacks;
c) comparing the outputs from operation b) for the subset of the plurality of packages in the one of the plurality of stacks to a list of a plurality of expected SKUs that are expected to be in the one of the plurality of stacks, wherein the list of the plurality of expected SKUs in operation c) is a list of SKUs to be shipped to a store;
d) repeating steps a) to c) for each of the plurality of stacks of the plurality of packages;
e) based upon the at least one machine learning model, forming a plurality of groups of images of the plurality of packages in the plurality of stacks; and
f) based upon the lists of the expected SKUs, assigning one of the plurality of expected SKUs to each of the plurality of groups of images.
24. The system of claim 23 further including the operation of: prior to step a) presenting the list of the plurality of expected SKUs to instruct a user to place the subset of the plurality of packages in the one of the plurality of stacks.