AI.SEE™ – Automated Quality Assurance
using AI driven Image Recognition

AI.SEE™️ is a visual quality control and defect detection system for manufacturers based on artificial intelligence. It consists of the self-learning software solution AI.SEE™️ Control and the hardware components AI.SEE™️ Lens, a highly flexible camera and lighting system, and AI.SEE™️ Edge, which evaluates captured images and controls the subsequent process.

Plug’n’Play

Easy integration by using a standardized camera and lighting system

Intelligent

Self-learning error detection through neural networks

Manufacturing 4.0

Centralized data networking and monitoring of the entire production quality

Challenges in Quality Assurance

The demands put on modern production facilities are tremendous: They are expected to be highly productive, flexible at maximum quality, cost-efficient, and trouble-free.

Manual and other conventional quality inspections are usually unreliable, expensive, have a low detection rate and are difficult to scale.

By taking advantage of the possibilities offered by artificial intelligence, AI.SEE™️ enables producers to perform a non-destructive, reliable, and automated inspection of their products.

With deep-learning algorithms, you can find even the smallest defects on heterogeneous or reflective surfaces that human inspectors often overlook and deliver potent insights while maintaining and improving reliable quality and operational efficiency.

Features

Annotation and Management of Image Data

With AI.SEE™️ Control, individual images can be inspected in detail, annotated, and assigned to defect classes. Furthermore, images can be managed in sets and assigned to projects to use them for training and retraining of different deep learning models.

Management of Projects and User Rights

Individual projects can be managed and analyzed within the AI.SEE™️ project database.

Also, personalized access to functions, projects, and elements within AI.SEE™️ can be defined for users.

Training, Deployment, and Evaluation of Neural Networks

AI.SEE™️ trains neural networks for error detection in the background. Should defects or components change, AI.SEE™️ Control can adapt to it, enabling a seamless transition between different product groups. Neural networks are trained, deployed, and evaluated within the system.

Dashboard and Data Analysis

Images, failure classes, and training data can be managed in AI.SEE™️, and the results can be examined on well-structured dashboards. For example, statistics and trends can indicate wear and tear of production machines at an early stage.

Interfaces for Different Imaging Systems

With AI.SEE™️ Lens, elunic provides its own camera and lighting system and enables a seamless integration to existing systems.

Interface for Deployment of Own Architectures and Models

Due to the rapid progress of machine learning technologies, any neural network architecture can be used. Their deployment can be achieved by the simple implementation of the AI.SEE™️ Engine interface.

Interfaces for System Integration in MES/QMS/…

To enable continuous production in a large factory, a precise allocation to each production piece is essential. This also includes marking if such production items do not meet the quality requirements.

AI.SEE™️ Benefits

Cloud & On-Premises

Can be run on modern cloud platforms or locally in your factory

Traceable

Transparency and traceability of product quality in automatic reports

Scalable

Suitable for use in numerous industries and production environments

Flexible

Individual camera, lens, and lighting setup

Easy to integrate

Simple installation and setup in existing systems and production facilities

Highly Efficient

Automated identification of various defect types within seconds

Can Artificial Intelligence Automate your Optical Inspection?

When it comes to supporting companies in manufacturing, similar challenges and concerns frequently arise. To provide you with a proper evaluation of the feasibility and price, please fill out the questionnaire below.

AI.SEE™️ Presentation at the IAS in Shanghai

``Data is the new gold. Image data can be managed and annotated in the central data hub. The setup takes less than 30 minutes on average!``

Rule-Based Computer Vision vs. Deep Learning

Rule-based computer vision used in conventional AOI test systems only deals with identifying images based on fixed rules for the recognition of lines and shapes.

In contrast, AI.SEE™️ uses machine learning models that learn the interrelationships between diverse, even the most complex data sets in real-time using state-of-the-art artificial neural networks through deep learning. These enhanced adaptabilities not only provide superior quality but also improve the performance of production equipment.

AOI (automatic optical inspection) enables fast and accurate automated inspection of electronic assemblies, e.g., to ensure high product quality in PCB manufacturing. Surface inspection, position, orientation sensors of electrical components, and classifications are only a few examples of the system’s application possibilities. Unlike conventional image processing systems, AI.SEE™️ uses state-of-the-art technologies to detect and locate objects and defects more efficiently and improves the throughput and ease of use.

``For producers and machine manufacturers, AI.SEE™️ is the ideal solution on your way to connected factory and Industry 4.0.``

4 Steps towards AI-supported Quality Assurance

Step 1:
Proof of Concept

1 defect class on 1 material/product

Start straight away and validate your AI-based quality assurance project fast, cost-efficient, and convenient with an online showcase to detect a defect class on a material or product. Our experts are ready to help you find the right camera system for your solution or to install our own camera and lighting solution—AI.SEE™️ Lens.

Camera system: If you already have an existing camera system installed, we can dock onto it without any problems.

Step 2:
Comprehensive Online Use Case

All defect classes on all materials/products

In the second step, the project will be extended. The deep learning model will be further trained to detect all defect classes on all materials and products in your production and to directly display and analyze the different defect classes on newly uploaded images.

Step 3:
Autonomous Installation in Production

All defect classes on all materials/products

After training your AI.SEE™️ project for all defect classes across all materials or products, you will receive your ready-for-use detection solution on AI.SEE™️ Edge, our performance computing system optimized for artificial intelligence, to integrate image analysis into your production.

Step 4:
Retraining and Remote Quality Assurance

All defect classes on all materials/products

To ensure a high detection rate, AI.SEE™️ launches into production with high error sensitivity. False positives are marked as such during operation to reduce the number of error messages on error-free images (false positives). With this data, AI.SEE™️ continues to learn automatically, reducing the number of false positives daily.


Remote Quality Assurance:
Enable your staff to access quality control results and make annotations regardless of where they are located.

``A camera or just a mounted smartphone is all you need to start with AI-supported quality assurance!``

AI.SEE™️ Products

AI.SEE™ Lens

AI.SEE™️ Lens is a versatile camera and illumination system that contains an optical lens and integrated lighting, specially optimized for surface inspections within industrial production processes.

AI.SEE™ Edge

AI.SEE™️ Edge analyzes captured images locally and controls further processing depending on the result, e.g., automatic separation of defective products or displaying warning messages.

AI.SEE™ Control

AI.SEE™️ Control is a cloud-based portal on which the gathered image data can be annotated to develop the self-learning A.I. behind it continuously.

``AI.SEE™️ runs in the cloud or on your shop floor. And you can test the software for one month for free.``

Referenzen

Contact.

Do you have a project inquiry or just want to say hello? Please do not hesitate to contact us – we will be happy to answer you immediately.

    *By using this form, you agree to the storage and processing of your data on this website.

     

    elunic AG

    Erika-Mann-Str. 23

    80636 München

    Deutschland

    089 / 4161737 - 30

    info@elunic.com