Fizyr specializes in segmenting, classifying, quality control and finding the best grasp poses of parcels that vary in shape, size, color, stacking, material, etc. Our deep learning algorithms are used to enable automated picking of unknown parcels from bulk for e-commerce, warehousing, and parcel handlers.
Fizyr’s modular software enables robots to cope with unknown parcels in parcel picking and placing. Our algorithms successfully allow robots to:
- pick parcels directly from conveyed bulk
- infeed items to induction lines of sorters (parcel induction)
- singulate closely stacked or overlapping items (ex. white-on-white / black-on-black flats, envelopes, flyers, deformable objects, etc.)
- detect labels facing up or not
- accurate (inclined) placement of parcels to feed sorters
For an extensive list of more objects Fizyr algorithms can detect, visit our solution page.