Fizyr has strong experience in helping worldwide leading integrators build the best, robust, and cost-effective picking cells.
World class algorithms
Fizyr applies deep learning to enable automated picking of unknown items for years. Our algorithms provide over 100 grasp poses each second, including classification to handle objects differently, but also perform quality controls and detect defects to prevent damaged items from being placed on a sorter.
Fizyr algorithms provide all relevant information, including:
- Segmentation (incl. white on white flats or reflective bags)
- Classifiers of type of parcel (incl. box, bag, envelope/flat, tube, cylinder, deformable, etc.)
- Determining outliers or non-conveyables
(i.e. damaged goods)
- Best possible grasp poses in 6 DoF
- Multiple ordered poses per object
Robust object detection in various conditions
Fizyr’s computer vision software successfully detects a large variety of objects for multiple logistics applications. Using a proper stereo camera for triangulation, our algorithms properly segment, classify, and propose the best possible grasp poses.
Trained with a unique dataset with millions of images of challenging logistics environments, our algorithms can also provide valuable information on damages, package material and slip sheets which are relevant at picking cells in harsh logistics environments.
Detection list including:
- Free-hanging straps
- Cardboard boxes/bags
- Deformable objects
- Loose plastic
- Jiffy/smart/flyer bags
- Transparent objects
- Bags of milk powder
- Damaged goods (incl. gradation)
- Envelopes / Flats Reflective packaging
- Beauty products
- Dangerous goods (ex. flammables)
- White-on-white / Black-on-blackApparel/clothing in polybags
- Closely stacked objects
- Fragile items
- Tubes / Cylinders
- ElectronicsOverlapping objects
- Loose loadSatchels
- Trays Label (General/Handler-specific)
Easy and fast integration
Fizyr has created a standard “Plug and Play” modular product that allows for a smooth integration with any system. We help our integrators configure the best picking cell, while facilitating cost, robustness, and speed optimization.
With an easily configurable interface, Fizyr’s software runs on premises, in a standard computer, with no need for internet access. It is also hardware agnostic, which gives our clients the freedom to choose the best camera, robot and gripper for their automated picking cell and remain flexible in changing their equipment any time they want.
FIZYR IS UNIQUE
We successfully grew bootstrapped delivering value for money.
We are the experts and know how to create true value for our partners.
We have the interfaces to manage numerous installations for updates and upgrades and give clients full control over our software through unique tools.
Through easy-to-understand specialized agreements, we make sure our clients continue to succeed also in case something happens to us (escrow).
Let’s discuss how we can help you.
6 DoF Grasp Poses Proposing the best grasp strategy based on all classifiers. You can configure your preferred…
From boxes and trays to deformable objects, from closely stacked or overlapping objects to damaged goods, Fizyr…
Automated Picking for Robotics Automation
Enabling robots to pick unknown objects from bulk Computer vision software for automated picking Fizyr is specialized…
Trained and proven for eCommerce and groceries Trained with a unique dataset of more than 1.5 million images of…
Enabling robots to grasp unknown parcels from (conveyed) bulk Plug-and-play vision software for parcel handling Until…
Fizyr's modular software segments, classifies, does quality control and finds the best grasp poses of items and parcels…
Originally published at https://fizyr.com.