KOMPASS brings lot segregation to the line, classifying every item by batch or lot as it moves. AI vision reads lot identity from codes, labels or visual markers, confirms each item belongs to the batch being packed and routes or flags any item from another lot.
Traceability depends on each pack holding items from one known batch. Without lot segregation on the line, lots mix at changeovers, merge points or rework, and a recall has to include every lot that might be inside.
Our AI Product Inspection system, KOMPASS™, enforces lot segregation by reading each item's lot identity from codes, labels or visual markers and checking it against the batch being run. Each lot decision is stored with an image of the item.
Lot segregation depends on reading each item's identity reliably and knowing which batch should be running. KOMPASS combines both with a direct link to diverters and your production records.
A camera captures every item where its code, label or marker is visible, with lighting set for the surface.
The model reads the lot identity and compares it with the active batch, or assigns the item to the right lane.
Items from another lot are routed or flagged, lot boundaries are logged and every result is stored.
Send us examples of how lots are marked and where batches can mix. We will assess the view, reading and routing your lot segregation needs.
Get In TouchEach method supports traceability. The difference is whether lot segregation checks every item where batches can mix.
Lot segregation with KOMPASS means reading the lot identity of every item on the line and confirming it belongs to the batch being packed. Items from another lot are flagged or routed to a separate lane. The result is a line where batches stay apart automatically, instead of relying on operators to keep lots separate through procedures and paperwork alone. Each decision is stored with its image, which gives quality teams evidence for every pack.
Most often from lot codes printed on the item, its label or its pack, read with AI OCR trained on your print. Where items carry no readable code, visual markers can be used instead, such as colour tags, label variants or tray markings. The method depends on how your process marks each batch at the point where items can mix. Items whose lot cannot be read are sent for review rather than assumed.
Mixed batches usually appear at merge points, accumulation tables and buffers, where items from the end of one run meet the start of the next. Rework, returns and manual top-ups are other common sources. A check placed after these points catches stray items before they reach packing, where mixing becomes far harder to undo. Placing the check at the right point keeps a small mix-up from becoming a large one.
Each item's lot is recorded automatically with its image and time stamp. That strengthens lot traceability by proving which batch went into which pack, without relying on manual logs that suffer transcription errors. If a recall is needed, a clean record narrows the recall scope to the lots actually affected. It also saves time during audits, because the evidence is already organised by lot.
Yes. During a batch changeover, the station records exactly when the last item of the old lot and the first item of the new lot passed. Those lot boundaries are stored automatically, so production records show where each batch starts and ends. Items that appear out of sequence after the change are flagged immediately. The boundary record can be shared with your production system for complete batch documentation.
Where several batches run together by design, divert lanes can separate items by lot after the camera classifies them. Each item is tracked to the diverter so it reaches the right lane. Where only one batch should be running, a stray item is rejected or the line is stopped instead. Lane assignments are logged so the contents of each lane can be confirmed later.
Verifying a lot code checks that a printed code matches what should have been printed. Lot classification uses the code, or another marker, to decide which batch an item belongs to and to keep batches apart. The two work together: one confirms the code is right, the other uses it to protect batch identity. Running both checks at one station is common where codes are printed just before packing.
Send examples of how lots are marked on your items, where batches can mix on your line and how batch changes are managed today. Add the line speed and where a reject or divert lane can go. With that, we can recommend camera placement, reading and routing for lot segregation and a clean audit trail on your line. Items from both sides of a recent batch change show the station exactly where lots meet.
Share how lots are marked, where batches meet and your line speed. We will plan lot segregation at your merge points and give you timing and cost. KOMPASS projects of this kind usually finish in 6 to 8 weeks.
