Zero miscounts
Vision-based counting eliminates the shortfalls and overages that manual tally and weight-based estimates let through.
Accurate Part Counts, Every Batch
Counting that never loses track
Whether parts are dropping through a chute, moving along a belt, or already sitting in a tray or stack, vision-based counting catches the shortfalls and overages that manual tally and weight-based estimates miss — at full accuracy.
Why counting goes wrong
Manual counts and weight estimates both drift
A missed capsule in a bottle, a short stack of rings before assembly, an SKU count that doesn't match the dispatch manifest — these are quiet, expensive errors. Manual counting is slow and fatigues; weight-based estimation breaks down the moment part weight varies. Vision counts the actual units, every time, and logs the result.
Three ways parts are counted
Pick the setup that matches your line
Free-Fall Counting
Parts dropping through a chute, spout, or funnel — capsules, tablets, fasteners, kernels — counted as they fall, at high speed.
ExploreConveyor-Based Counting
Parts moving continuously along a belt or line — bottles, pipes, packaged goods — counted in motion, encoder-synced to line speed.
ExploreStatic Tray / Stack Counting
Parts already arranged and stationary — a stack of rings, a tray of vials, a pallet face — counted in a single capture.
ExploreCapabilities
What this solution delivers
Full line speed
Counts every unit inline — no sampling, no slowing the line down to check a batch.
Auto reject on mismatch
Flags or diverts any pack, stack, or batch that doesn’t match the required count before it moves downstream.
Typical applications
Where it's used
Capsules & tablets
Fasteners & small hardware
Rings & stacked components
Vials & ampoules in trays
Pipes, bars & packaged goods
Pallet / SKU face counts
Proven in production
Case studies & app notes
Case studies
AUTOMOTIVE INDUSTRIESCASE STUDYCounting Piston Rings with 100% Accuracy Using Machine Vision
Jun 17, 2021READ →
PHARMACASE STUDYVial Counting Using Image Processing to Improve Inspection Accuracy and Reduce Downtime
May 5, 2022READ →
LOGISTICS AND PACKINGCASE STUDYCount the SKUs in a Pallet
Aug 12, 2021READ →
App notes
FMCGAPP NOTEMulti-Lane Biscuit Counting with One Camera
A single camera above the post-oven conveyor replaces per-lane photoelectric sensors, using instance segmentation to count biscuits, flag broken or deformed pieces, and track lane balance across 2–3 lanes — addressing the specific failure mode that causes beam sensors to undercount once biscuits touch or shingle.
READ →
FMCGAPP NOTEBottle Cap Counting with In-Line Defect Detection
A single overhead AI camera counts bottles and screens caps for missing, skewed, wrong-colour, and cracked defects in one inference pass — building the kind of timestamped, batch-linked image record that FSSAI's revised packaged-water testing regime is moving toward, instead of relying on a torque sensor and a fatigued spot-check.
READ →
INDUSTRYAPP NOTEAutomated Part Counting for Mixed-SKU Electronics Using Machine Vision and Deep Learning
An instance-segmentation machine vision architecture for counting mixed-SKU electronics and interconnect parts — from sub-5 mm micro-pins to 200 mm cable assemblies — in a single top-down image, built to remove the fatigue-driven error floor that published human-factors research documents in manual counting.
READ →
FMCGAPP NOTEDense-Pack Tray Counting for Tobacco Manufacturing Traceability
Vision-based count verification and SKU traceability for cigarette buds — replacing an assumed tray-fill count with a measured, image-backed, and audit-ready result at every tray, built around the count-tolerance and cycle-time targets a pilot actually validates rather than a number quoted off a spec sheet.
READ →
Proven in production
Case studies & app notes
Case studies
AUTOMOTIVE INDUSTRIESCASE STUDYCounting Piston Rings with 100% Accuracy Using Machine Vision
Jun 17, 2021READ →
PHARMACASE STUDYVial Counting Using Image Processing to Improve Inspection Accuracy and Reduce Downtime
May 5, 2022READ →
LOGISTICS AND PACKINGCASE STUDYCount the SKUs in a Pallet
Aug 12, 2021READ →
App notes
FMCGAPP NOTEMulti-Lane Biscuit Counting with One Camera
A single camera above the post-oven conveyor replaces per-lane photoelectric sensors, using instance segmentation to count biscuits, flag broken or deformed pieces, and track lane balance across 2–3 lanes — addressing the specific failure mode that causes beam sensors to undercount once biscuits touch or shingle.
READ →
FMCGAPP NOTEBottle Cap Counting with In-Line Defect Detection
A single overhead AI camera counts bottles and screens caps for missing, skewed, wrong-colour, and cracked defects in one inference pass — building the kind of timestamped, batch-linked image record that FSSAI's revised packaged-water testing regime is moving toward, instead of relying on a torque sensor and a fatigued spot-check.
READ →
INDUSTRYAPP NOTEAutomated Part Counting for Mixed-SKU Electronics Using Machine Vision and Deep Learning
An instance-segmentation machine vision architecture for counting mixed-SKU electronics and interconnect parts — from sub-5 mm micro-pins to 200 mm cable assemblies — in a single top-down image, built to remove the fatigue-driven error floor that published human-factors research documents in manual counting.
READ →
FMCGAPP NOTEDense-Pack Tray Counting for Tobacco Manufacturing Traceability
Vision-based count verification and SKU traceability for cigarette buds — replacing an assumed tray-fill count with a measured, image-backed, and audit-ready result at every tray, built around the count-tolerance and cycle-time targets a pilot actually validates rather than a number quoted off a spec sheet.
READ →
Not sure which counting setup fits your line?
Tell us how your parts move — free-fall, belt, or something else — and we'll scope the right approach.