March Code
Service

Custom Computer Vision for Manufacturing

We catch the defects a tired inspector misses: flaws, unit counts and markings at conveyor speed. Send 100–300 photos from your line, and in 3–5 days we'll show you the accuracy on your products for free. Counting or OCR with one camera from $14,900, a defect detection pilot on one line from $19,900 in 2 weeks. Edge processing on the shop floor, and the line keeps running

Demo in 3–5 daysfree, before any contract
from $14,900transparent quote
13+ yearsof founder experience · US, Europe, CIS

Free consultation · 30 minutes · No obligation

Cost
from $14,900
First step
A demo on your photos in 3–5 days, free
First paid stage
Product counting or marking OCR: from $14,900
  • Quote in 24 hours
  • Pay by stage
  • Your code from day one
  • Weekly demos
  • 3-month warranty
  • Contract and invoices

Telegram @marchcode · info@march-code.tech · Mon–Fri 9:00–18:00 Central European time, and on Telegram we reply outside these hours too

When it matters

Situations where we help

If you recognize even one, we can be useful

01

Inspectors get tired and miss defects

By the end of a shift, attention drops, and some defects reach the customer. One missed defect on a critical part means a customer claim and a penalty. Computer vision checks every unit with the same accuracy at the start and the end of the shift, with no fatigue and no human factor.

02

100% inspection slows the conveyor

Checking every item by hand means slowing the line, so inspection is often done by sampling, and defects slip into the batch. A machine vision camera inspects the flow at full conveyor speed and rejects defects in real time without cutting throughput.

03

Product counts and markings are tracked by hand

An operator counts items, checks batch numbers and markings by eye and copies them into a logbook, which leads to accounting errors and mixed-up stock. The system counts items automatically, reads markings with OCR and sends the data to your accounting system with no manual entry.

04

The rejection threshold is an endless argument between shifts

One inspector rejects everything, another lets borderline cases through, and the defect rate jumps from shift to shift. The model holds one threshold around the clock, and operator corrections build up in the dataset and retrain the system: the quality standard stops depending on who's on shift.

Let's estimate the budget for your task

Describe your task and we'll come back with a timeline and cost estimate within 1 business day. Not ready for a quote? Send 100–300 photos from your line, and in 3–5 days we'll show the accuracy on your products, free.

Get a quote
What you get

Features and benefits

Concrete results we deliver for you

01

Defect detection on the conveyor

The system finds chips, cracks, scratches, paint gaps and shape deviations at line speed, rejects defective items in real time and records the defect type. It catches the defects that become hard to see by the end of a shift.

02

Product counting and tracking

The camera counts items in the flow, on a pallet or in packaging with no manual recount, and the data goes to your ERP (SAP Business One, NetSuite and the like) or MES automatically. In the ProControl case, that's 200,000+ pallet scans and 100% of output tracked with no manual entry.

03

Marking and number recognition (OCR)

The system reads batch numbers, serial numbers, dates and barcodes right off the product, checks them against the job and flags incorrect markings. It works with print, engraving and labels in shop floor conditions.

04

Defect statistics by shift and line, in your ERP/MES

Every scan is logged: defect type, line, shift, time. Instead of a notebook and gut feeling, you get statistics that show where and when defects originate. ERP and MES integration out of the box.

05

Works with your cameras and your data

We train the model on your products and on defect samples from your line, so it knows your defects specifically. We connect to existing cameras or add optics and lighting only where needed, without replacing equipment.

06

On-site deployment, no cloud

The system runs on an edge device right on the shop floor: data never leaves the plant, and there's no dependence on the internet. Our ProControl scanner is a Raspberry Pi with a 3D camera in an industrial enclosure: dust, vibration and temperature swings don't bother it.

Process

How we work

A transparent process with clear stages and deliverables

1
Stage 1

Demo on your photos: 3–5 days, free

Send 100–300 images of good and defective items; phone photos or an export from your camera will do. We train a model on one of your defects and show its accuracy on your own sample. We don't connect to the line, and you keep the report, before any contract.

2
Stage 2

Pilot on one line: 2 weeks, from $19,900

A camera with lighting at one inspection point, an edge device on the shop floor, and a measurement of accuracy and false-positive rate on the real flow. Target metrics are fixed in the contract before we start.

3
Stage 3

Scaling the model: 2–4 weeks

We label the full dataset and retrain the model for all defect types and shift conditions. We test on a held-out sample until we reach the target accuracy and a low false-positive rate.

4
Stage 4

Installation and integration: 1–3 weeks

We install cameras and lighting without stopping the line; installation takes hours, as it did at ProControl. We connect to the PLC for automatic rejection and to your ERP or MES for statistics. We calibrate for real production conditions.

5
Stage 5

Launch and retraining: ongoing

We track accuracy and the defect rate at the output. We retrain the model on new defect types and seasonal changes in raw materials, from $1,300/month. SLA on-call duty is a separate software support subscription.

Cost and timeline

What it costs

You get an exact quote within 24 hours of the call

Cost
from $14,900
What's included
  • Start free: a demo on your photos in 3–5 days. Send 100–300 frames of good and defective items, and we show the model's accuracy on your sample. We do the labeling, we don't connect to the line, and you keep the report
  • Pilot on one line or one defect type: from $19,900, 2 weeks. A camera with lighting, an edge device on the shop floor, accuracy measured on the real flow. Target metrics are fixed in the contract
  • Product counting or marking OCR with one camera: from $14,900. Simpler than defect detection, often launched first, and pays off faster
  • A quality control system for all defect types, with camera installation, PLC integration for automatic rejection and ERP/MES integration: from $34,000 in stages, 2–4 months
  • No enterprise-grade hardware: the ProControl scanner is a Raspberry Pi with a 3D camera in an industrial enclosure, installed in hours without stopping the line. The dataset, model weights and code are yours. Retraining from $1,300/month

The final cost depends on scope. We'll work it out after a short call.

Prices shown are indicative and for information only; they are not a binding offer. The final cost is set in the contract once the project requirements are agreed.

Computer vision without a corporate budget

Camera-based quality control is no longer reserved for giant factories. We start for free: send 100–300 photos of good and defective items, and in 3–5 days we'll show the model's accuracy on your sample, before any contract and without connecting to the line. Next comes a pilot at one inspection point from $19,900 in 2 weeks, and a full quality control system from $34,000 in stages; all price ranges are on the pricing page. How it works and what it gives a plant, we explain in “Computer vision in manufacturing”.

Proof from the shop floor, not from a slide deck: ProControl, a paving slab plant with four lines, 200,000+ pallet scans, 100% of output tracked with no manual entry, losses cut by up to 15%. Our own scanners are built on a Raspberry Pi with a 3D camera in an industrial enclosure and installed in a few hours without stopping production. Output statistics flow into the accounting system, and we connect to your ERP and MES through system integration.

Where the line is drawn: counting visitors and monitoring queues, PPE and the perimeter is video analytics, from $7,500 for a PoC. Quality control of products at line speed with automatic rejection is what this page covers, and it calls for a different level of accuracy. An industry overview for manufacturers is on our manufacturing industry page.

Case studies

Real results

Projects where we applied this service

Case study: Computer vision system for managing paving tile production
200K+pallet scans processed by AI

Computer vision system for managing paving tile production

Manufacturing

A manufacturing execution system (MES) for production control and output tracking at a paving tile plant. Our own scanners on the lines photograph every pallet, a neural network counts tiles and defects, and planning and analytics update in real time.

200K+pallet scans processed by AI
100%output tracking with no manual input
14 msAI tile detection per scan
4 linesrunning our own scanners
FAQ

Frequently asked questions about Computer Vision

Answers to the questions we hear most often

A demo on photos from your production line in 3–5 days, free

Demo free · counting or OCR with one camera from $14,900 · defect detection from $19,900

Prices are indicative and not a binding offer.

Send 100–300 photos of good and defective items. We train a model on one of your defects and show its accuracy on your own sample. We don't connect to the line or stop production. You keep the report, before any contract or prepayment.

Step 1 of 2 · Task

Send product photos on Telegram, no form

Keep exploring

Read next

AI IntegrationService

AI Integration

We bring artificial intelligence into your business processes: chatbots and AI agents, computer vision, predictive analytics, digital twins. From $3,900, with a free demo on your process in 5 days.

Learn more
IntegrationsService

Integrations

We connect the systems you already run: your ERP or accounting system, CRM, website, phone system and marketplaces exchange data on their own, and managers stop re-typing orders and stock by hand. One integration from $4,900 in 2–3 weeks, a free integration map in 3–5 days. We don't touch your ERP's core.

Learn more
Video analyticsService

Video analytics

Video analytics on your existing cameras: visitor counting and heatmaps in retail, PPE and workplace safety monitoring in production, restricted zones and perimeters, license plate recognition at the gate. A demo on footage from your cameras is free, takes 3–5 days and comes before any contract. A PoC on 1–2 cameras from $7,500 in 2 weeks.

Learn more
Computer Vision in Manufacturing: How It Works and What It CostsArticle

Artificial intelligence · 15 min

Computer Vision in Manufacturing: How It Works and What It Costs

How machine vision checks quality, counts products and catches defects on the production line. The technology, implementation costs and a real case with numbers.

Learn more
Video Analytics for Business: Visitor Counting, Quality Control, SecurityArticle

Artificial intelligence · 15 min

Video Analytics for Business: Visitor Counting, Quality Control, Security

How video analytics counts visitors, monitors service quality and improves security. PoC from $7,500, a free demo on your own footage, works with your existing cameras, real cases.

Learn more

Manufacturing

Computer vision system for managing paving tile production

A manufacturing execution system (MES) for production control and output tracking at a paving tile plant. Our own scanners on the lines photograph every pallet, a neural network counts tiles and defects, and planning and analytics update in real time.

Learn more