Soldering quality control prototype
Company
OSRAM a.s.
Project Duration
1 month
Implemented System
Visual inspection system
Area
Manufacturing quality control
Our customer
Our company was approached by a client to improve an existing vision system designed for soldering inspection. The original solution was implemented as a classic OCR-based system (object recognition), aimed at identifying incorrect soldering and subsequently sorting out defective products. However, the original system produced a high number of false positives, and when tolerance levels were reduced, the solution failed to meet the required quality standards. Our project focused on research and development of a new solution that would ensure compliance with strict quality standards through the implementation of an artificial intelligence–based system.
Our challenge and objectives
The project team designed a new solution based on the use of a camera system and deep neural networks. The objective was to identify algorithms and approaches that would significantly increase accuracy during product inspection.
The goal was to develop an automated visual quality inspection system, capable of reliably detecting defects using deep neural networks and increasing the overall efficiency of the production process.
Our solution
The developed quality control prototype enabled fast and efficient defectoscopy of soldered light bulbs.
The system is capable of continuous operation and achieves reliable accuracy exceeding 90%. The solution we developed improves the quality of the final product and ensures that soldered light bulbs meet the high quality standards required in the industry.
Within the project, we focused on the use of deep neural networks for identifying and controlling the quality of soldering on light bulbs.
Benefits for the client
Accurate quality control
A system based on deep neural networks reliably detects defects and errors on each light bulb with accuracy above 90%.
Increased roduction efficiency
Automated processing of camera images accelerates the entire inspection process and minimizes the need for manual intervention.
Use of advanced know-how
Our solution introduced techniques that are now used on an automated production line in Germany for inspecting solder joints on light bulbs.
Project complexity
Size
2/6
Time
1/6
Finances
2/6
Complexity
6/6
Are you interested in a similar solution?
We will be happy to help you create a system tailored to your needs – from design to implementation.