Core Sorting AI for Remanufacturing
Identify every returned core by part number, even when the label is gone, then grade its condition and route it to remanufacture, credit, or scrap at the receiving dock.
Core Sorting AI for the Toughest Engine, Drivetrain, Electrical, and Brake Remanufacturing Lines
Deploy Anywhere, Run Everywhere
Run core sorting on receiving dock, conveyor, and sorting table cameras on the edge, on-prem, in your VPC, or via API.
One Platform, Full Adoption
Tools every remanufacturing team can adopt, from core receivers and sorters to reman engineers and returns teams, no separate ML team required.
Secure, Compliant, and Audit-Ready
Data stays safe with SOC 2 Type II compliance, encrypted data, HIPAA compliance, and an uptime SLA, with an image behind every core ID and credit decision.
- Core Identification by Part Number
- Look-Alike Variant Disambiguation
- Worn & Painted-Over Marking OCR
- Condition & Damage Grading
- Missing & Non-Original Components
- Reman, Credit & Scrap Routing
“Roboflow has been instrumental in accelerating our deployment of innovative AI solutions.”
Travis Turnbull
Vice President & CIO, Pella Corporation
Talk to a Vision AI engineer who's shipped in automotive and remanufacturing.
Bring us your toughest core sorting problem and we'll map a working solution.
Ask us about:
- Solution architecture that fits your part catalog, acceptance rules, and ERP and core credit systems
- A live demo on your own core receiving camera footage
- Deployment options: edge at the dock, on-prem, air-gapped, VPC, or on the cameras you already run
- ROI modeling against scrapped good cores, misidentified credits, sorter hours, and teardown rework
Over 16,000 organizations build with Roboflow.
- Rivian
- Pella
- Chobani
- USG Corporation
- BNSF Railway
- American Woodmark
Vision AI is transforming businesses
Customers across the board are solving complex challenges and driving meaningful impact.
- Automotive customer
$10 million
Saved by automatically detecting defects on the production line
- Logistics & freight company
90%
Less time spent manually tracking shipping inventory
- Building materials supplier
60%
Lower customer return rate with improved product quality
Identify Every Core, from the Receiving Dock to the Teardown Line
Add a real-time identification layer to core receiving with Roboflow vision AI for core sorting. Built for remanufacturers where returned cores arrive dirty, rusted, and missing their labels, two variants with different part numbers can look almost identical, and a core that can't be identified gets scrapped instead of remanufactured.
Identify the core:
- Recognize the part type and variant from shape, size, and features such as mounting points, connectors, and housings, with a model trained on your own returned cores.
- Tell apart look-alike variants that carry different part numbers, so a core is credited and routed against the right part.
- Read cast, stamped, and data plate part numbers with industrial OCR, including worn, dirty, and painted-over marks and laser-etched codes.
Grade the core:
- Find cracked or broken housings and castings, heavy corrosion, and fire or impact damage that make a core non-remanufacturable.
- Detect missing components and non-original parts, and grade wear on friction surfaces such as brake shoe linings.
- Apply your own acceptance rules to each part family, so every sorter grades the same core the same way.
Route and credit:
- Route each core to remanufacture, hold, or scrap, and send it to the right bin, rack, or teardown line.
- Attach images to every core credit and return decision, so disputes are settled with a photo instead of a phone call.
- Send low-confidence IDs to a sorter to confirm, and feed every confirmation back into training, the same approach as part identification and scanning.
Bring intelligence to every core today. Stop good cores from going to scrap because nobody could tell what they were.
Frequently asked questions
What is core sorting with Vision AI?
Core sorting with Vision AI uses cameras at core receiving and a trained computer vision model to identify each returned core (the used part sent back for remanufacturing), read its markings, grade its condition, and route it to remanufacture, credit, or scrap. Every decision is saved with its image and the part number it was matched to.
Can it identify cores with missing or unreadable part numbers?
Yes. The model identifies a core from its appearance, such as shape, size, and features, and uses OCR on any markings it can read. When it is not confident, the core goes to a sorter to confirm, and that confirmation becomes training data, so the model gets better at the variants your dock sees most.
Can it handle thousands of part numbers and new variants?
Yes. Most teams start with their highest-volume part families, then add variants as confirmed sorts build up the training data. New part numbers can be added without rebuilding the system, and the same setup extends to new product lines.
Roboflow serves as the identification engine, and your remanufacturing and returns teams own the part catalog, acceptance rules, and every credit decision.
Can it integrate with our ERP, WMS, and core credit system?
Yes. Roboflow sends part numbers, grades, and images over REST, MQTT, or OPC UA, or writes them to a database, so they flow into ERP, WMS, and core credit and returns systems, and can drive a PLC that diverts each core on a conveyor. Models run on edge hardware at the dock, on-prem, or in your VPC.