Computer Vision


Vision System for Verification of Laser Welding

We have developed software for the motor manufacturing process, were approximately two hundred pins must be meticulously inspected, a task that can occasionally result in errors, posing safety concerns. 
In order to enhance accuracy and efficiency, a vision system has been integrated into a robot arm. The system employs innovative distance measurement techniques, relying on focus evaluation, in combination with Deep Learning algorithms.
This advanced technology ensures more reliable pin checking, improving safety and overall quality control within the manufacturing process.


Vision System for Quality Control

Software solution for quality control in the production of paint cans used by pharmaceutical and cosmetics manufacturers.

As part of the production of cans, the print must be applied legibly and without defects. The process takes only 2s. Every 2 sec. a new can is being sprayed and during this time our solution has been implemented whereby the can is being photographed 6 times with a high resolution camera and the images are being verified by the system. Even the most minute defects are being detected.
The search for defects is using DeepLearning type algorithms.


Vision System for Verification of Electronic Boards

Our team developed a solution for the verification system for soldering on electronic boards in a large LG battery factory.
Accurate soldering of components is crucial for safety reasons. Incorrect soldering can result in overvoltage, potentially leading to cell ignition and fires.
Adequate verification processes are imperative to mitigate these risks and guarantee the correct assembly and soldering of components to maintain safety standards.
Our solution included a camera and profilometer to enable 3D scanning and geometric verification. Advanced optics and lighting have been used to precisely identify small details.


Vision System for Verification of Operator’s Tasks

In the factory settings operators sometimes deviate from established procedures due to the tiring and monotonous nature of their tasks. To address this issue and enhance operational efficiency by verification of detail placement by an operator, a vision system was introduced.
We used Deep Learning vision system to enhance the verification process of the operator’s work. This technology plays a crucial role in supporting and augmenting the work of operators. It provides real-time assistance by monitoring and guiding operators through tasks, ensuring adherence to procedures. Moreover, the vision system aids in quality control by minimising human error, and enhancing safety.
Overall, this implementation significantly improves the overall productivity and accuracy of factory operations while mitigating the effects of fatigue and monotony on the workforce.


Automatic 3D scanning station

We develped an automatic 3D scanning station using professional blue light 3D scanner and a five-axis 3D scanning system that delivers models with scan quality in the range of single micrometers.
The platform incorporates a hardware module with a professional structure light 3D scanner, a desktop application that directly manages the work of the scanner and creates a solid model based on data obtained from the scanner, and a cloud platform for handling the scanning process, including ordering and managing scans and evaluating the resulting data.
Defective metal parts on oil rigs are beeing 3D scanned on the spot. Detailed scan is processed, sent to the 3D printer, and printed in metal on the spot without the need to ship broken parts on hundreds of
rigs from land.