Skip to content

Custom Vision Systems

Home » Custom Vision Systems

Vision Inspection Systems for Custom Needs

Unique challenges require unique builds. One of the benefits of the Artemis Vision process is that we work with customers in a collaborative, test-based approach from problem discovery through ongoing evaluation and support. Our solutions are built around the necessary test criteria and proven to work before they are delivered. We tailor software to meet specific needs of the client. Plus we provide the tools for any custom solution to be adaptable and scalable, with tools to track and diagnose quality problems, production data and trend reporting, and global uniformity of key parameters.

Browse some of our case studies below for ideas on starting from a semi-custom build, go to Our Turn-Key Manufacturing Solutions page, or jump straight to a quote request.

Rewind Table Label Verification System

Replace manual inspection with an automated system that detects smudges, misprints, and unwanted marks on every label, as well as verifying correct barcodes on labels.

This vision inspection system uses deep learning, pattern matching, and optical character recognition (OCR) algorithms to recognize defects, inconsistencies, and incorrect lot and expiration information. It scans and processes labels at a rate of one label per second. Alarms warn the operator of failed labels, which can be removed and recorded before continuing to inspect the remainder of the label roll. The HMI displays inspection data during the process, including the number of labels processed, number of failed labels, and the number that passed visual inspection. The system includes a configurable recipe editor, allowing customers to input different inspections for different label types.

Designed to work with a rewind table, this system enables you to seamlessly achieve 100% inspection while improving production speed and freeing up valuable personnel hours. Adhere to industry regulatory requirements, satisfy customer expectations, and improve your operational efficiency.

label inspection system on rewind table
failed barcode on label - vision inspection

A Challenge Resolved

A major medical company performed manual inspections on labels for pharmaceuticals and medical serums at a rate of about 13 labels per minute, individually checking six different fields on each label. With serious implications for incorrectly labeling items for patients and falling short of pharma regulations, as well as an interest in saving upwards of $50k a year in increased efficiency, they turned to Artemis Vision to develop an automated label verification system.

The system processes labels for improved quality and compliance. And it improved inspection speed over 5 times, from 13 to 60 labels per minute. With 600 dpi resolution and a 12-inch field of view, the system can now detect errors down to ~0.020. Valuable personnel hours have been freed up to use elsewhere, and inspection accuracy is assured and automatic.

detect missing and damaged labels

Bottle Label Inspection System

Helps you deliver safe and well-marked products to customers without having to manually check every label.

Our factory-tested and approved vision system uses a high-resolution camera, high-quality lens, and diffusion-axis lighting. Showcasing finely tuned deep learning performance, the system determines presence, orientation, position, and integrity of labels with a customizeable software to save multiple sets of tolerances for your own label needs. Authorized users can create distinct inspection recipes that are saveable so that the same recipe can be loaded for each time a label is run. A reject cylinder automatically pushes bottles from the line for all labels that do not meet specifications.

A Challenge Resolved

A major healthcare company had to implement costly and inefficient manual checks for scientific and medical serums. There was an unacceptable risk that bottles with missing, smudged, or damaged labels were unknowingly distributed. Artemis Vision’s machine vision system for label inspection automatically and reliably ensures that bottle labels are free of error.

We trained a neural network to recognize errors based on client samples, and then re-trained the neural network to repeatably detect smaller and smaller errors, to as little as 2 degrees of rotation and 1/10th of an inch in defect size.

This system was factory and site tested for bottles from 30ml-4L, and the resolution used was 4096 x 3000. It is configurable for label and product size.

label defects - smudge, wrinkle, misalignment

Container Cap Verification System

Inspects bottles and containers to ensure caps are sealed and threaded correctly and tightly, avoiding costly and time-consuming consequences.

A fully automated solution verifies caps are threaded correctly and foil seals are in place by a high-speed camera, lighting, lenses, and photo-eye. Easily integrated into existing production lines, the system is adaptable to high conveyor speeds and is floor-mounted to allow for installation over many conveyor types. We’ve ensured that the system is easy-to-use by equipping it with a quality operator HMI that is simple to install, easy to understand, and modifiable for user requirements. As bottles move through the conveyor, the throughbeam photo-eye, located on the leading edge of the bottle or container, triggers if it becomes occluded. A separate foil seal sensor indicates whether a foil is in place or missing. The program can complete a robust inspection for specified defects based on configurable pixel tolerances.

Easily deploy the system on current lines to add accountability, repeatability, and accuracy to the production process. With the vision system in place, manufacturers can now store images for historical analysis and can confidently answer questions and concerns from customers.

failed cap inspection
passed cap inspection
container cap inspection vision system

A Challenge Resolved

Pharmaceutical, food and beverage, and consumer good companies alike are facing quality and waste issues on production lines across the country. Tilted and skewed container caps are the result of incorrect cross-threading, which can lead to leaks, waste, and unnecessary costs. Manufacturers in these industries must meet strict quality standards and eliminate any errors associated with incorrectly sealed bottles before they leave the factory.

Artemis Vision worked with a local bottling plant to correct this frustrating problem. The Container Cap Verification System inspects bottles and containers to ensure caps are sealed correctly and tightly before leaving the facility. By eliminating faulty containers from the line prior to shipping, the manufacturer avoids excess waste and costs due to returns, leakage, and spoiled product. Capable of inspecting up to 8,000 bottles per hour, the system alerts operators of production issues as they occur, so they can be fixed before costly and time-consuming consequences.

angle detection for minting
coin angle detection system

Coin Die Alignment System

Identifies and corrects misalignment errors that operators and machine grinders tend to miss, meaning faster production times and a more confident manufacturing process.

This vision system met coining challenges by pairing with a client’s existing CNC Machine to assist in identifying accurate angles on the die before keying. There is only a small margin of error allowed for the stamping process, and the dies must be orientated correctly to ensure that there are no defects produced on the final product. The Artemis Vision system is capable of identifying and correcting those misalignment errors that the operator or grinder may not be successful in catching.

The unit illuminates the shiny die face and captures high-resolution images that are sent to the PC. Coin die angles are calculated via Artemis Vision’s proprietary software which then sends motion commands to the grinder to align the die. The system utilizes only industry-leading hardware for reliability, high dynamic range feature, and GigE interface capabilities. Additional hardware includes a machine vision lens, lens extender, and light diffuser.

Customizable features allow operators to set angular points on the coin die models with high accuracy for more consistent keying every time. The intuitive program also provides feedback on the images of expected angles, reported angles, and angle errors, and can retrieve previously saved models to save time.

A Challenge Resolved

Coin mints and manufacturers alike follow a detailed and highly accurate coining process to produce a high volume of new coins. The process involves using a CNC Machine Grinder to key coin dies that are used later during the stamping procedure. The CNC Machine must grind the die with high angular precision to ensure that it can be inserted with accuracy into the press. Mints can lose a lot of time, money, and efforts on consequent waste from improper coin stamping and must ultimately produce consistent coins with consistent quality.

Artemis Vision considered these production issues asked critical questions to determine the most successful vision solution.

A more detailed case study as originally published by Flir is found here.

Weld Inspection System

This efficient system with a small footprint automatically inspects welds to identify welding errors.

Cameras and a light-based system inspects the weld puddle from all sides on parts held in place by an actuator. Deep learning algorithms enable the system to recognize images of acceptable or failed welds. A low-profile red/green light instantaneously alerts for pass and fail and a light near the moving block alerts the operator if the part is inserted correctly.

A Challenge Resolved

A global medical technology company’s surgical equipment contains pins laser-welded in place during production. Personnel manually inspected the assembled devices, and occasionally the inspection was inconsistent. The company risked consumer health and safety and major liability issues in the event of any partial or missing welds. For greater quality assurance they turned to Artemis Vision for a system that repeatably safeguards against distributing any equipment with welding errors. The system is capable of inspecting varying lengths and different materials found in different versions of the devices. Every device is now inspected reliably every time, eliminating the worry of any missing welds, and improving patient outcomes and company reputation.

weld inspection pass fail
defect detection for heat exchanger plates
YouTube video

Micro-Hole Detection System

Speed up your production with automated, in-motion micro-hole detection on flat sheets of any size.

We combined a multi-camera line scan system, belt conveyor, and simple software to ensure excellent performance and an effortless experience for the operator. As material travels the conveyor, line scan cameras are encoder-triggered to image the surface for defects. The inspection system can detect hole defects down to .015″ in a 12ft sheet in under a minute.

A Challenge Resolved

Faulty heat exchanger plates lead to leakage and are a common HVAC headache. Inspecting heat exchangers by manual methods like dye testing is time-consuming and doesn’t guarantee 100% inspection. We developed our vision system for a company that was searching for a more efficient and non-destructive testing process. Their existing manual procedure required dye testing on parts that were already assembled. Our Micro-Hole Detection System integrated in-line for quick, 100% inspection, and enabled the manufacturers to catch bad parts pre-assembly. The high tech solution also yielded a much more accurate solution than the manual inspection process.

Glass Vial Inspection System

Automatically verify the integrity of molded glass vials with detailed precision.

glass vial defects deep learningThis vision defect detection system uses deep learning to confidently detect glass vial defects ranging from chipped finishes to seed bubbles. Designed for smooth transport without causing damage to the vials prior to, during, or after camera inspection, the system uses a starwheel to move the vials through imaging and carousel-based camera stations. The base system inspects at 120 vials per minute. Filter-matched lenses and lighting provide clear edge illumination. All inspection results and questionable images are stored in the system.


glass vial inspection system
sorting glass vials by defect
The system rejects defective vials and sorts them by defect.
YouTube video

A Challenge Resolved

A biopharmaceutical company uses glass vials to package an expensive medication, which must be discarded if any defects are found in the vial. Artemis Vision designed a system to identify vials with chipped, malformed, and line-over finishes, and even issues in the vial body like dope, seed, stone, and heavy mold lines. Built to ANSI and NEC standards, this vision system integrates inline to inspect moving glass vials prior to filling, eliminating the waste that arose from unknowingly filling damaged vials.

Contact Us for a Quote on a Custom Vision System

Please include any relevant documentation such as a User Requirements Specification or request for NDA form.

© 2024 Artemis Vision. All rights reserved. Privacy Policy.